PWiii iillipiil!!lliti!i^^

^'

f

ANNALS

SOUTH AFRICAN MUSEUM.

VOf.UMK /!'• ciintainiiiij : —

DESCRIPTIONS OF THE PALiEONTOLOGICAL MATERIAL

COLLECTED by the MEMBERS of the GEOLOGICAL SURVEY

OF CAPE COLONY.

PART I. nmtaininn :—^

1. Fossil Floras nf Ciijw Cnlnnn. With Foui-teeii Plates, and Eight
Text-figures. ]3y A. C. SSewakd, F.R.S., Univenity Lecturer
in Botavij, and Fellow of Emrnamiel Collec/e, Camhriihje.

if

ISSUED OCTOBEB 22, W03. PRICE Kh.

PRINTED FOR THE

TRUSTEES OF THE SOUTH AFRICAN MUSEUM

AND

THE GEOLOGICAL COMMISSION OF THE COLONY Ot THE
CAPE OF GOOD HOPE

By West, Newman & do., London.

ANNALS

OF THE

SOUTH AFKICAN MUSEUM.

I. — Fossil Floras of Cape Colony. — By A. C. Seward, F.E.S.,
University Lecturer in Botany and Fclloiu of Emmanuel College,

Camhridge.

Plates I.-XIV., Text-figures 1-8.

1.— FLORA OF THE UITENHAGE SERIES.

I.— INTEODUCTION.

The Uitenhage plants submitted to me for examination by the
Geological Commission of Cape Colony were collected from rocks
exposed in the valleys of the Bezuidenhout, Witte, Sunday, and
Zwartkop Elvers in the Uitenhage and Port Elizabeth districts in
the south-east of Cape Colony ; and from the neighbourhood of
Herbertsdale, Mossel Bay ; and Heidelberg.

In their ' Eeport on parts of the Uitenhage and Port Elizabeth
Divisions,' * Messrs. Eogers and Schwarz mention the following
localities from which fossils w^ere obtained : —

1. Zivartkop River, between the town of Uitenhage and the mouth
of the river.

2. Bezuidenliout River, from Blauw Krantz to the Sunday Eivev.

3. Witte River, from Enon to the Sunday Eiver.

4. Sunday River, between Dunbrodie and Addo Station.

* Eogers and Schwarz (01).
1

2 Annals of the South African Museum.

The Uitenhage series in the district of that name is described as
including the three following divisions in descending order : —

a. The Sunday Eiver marine beds containing animal fossils.

h. The Wood-bed with remains of fossil wood, ferns, and cycads.

c. The Enon conglomerate beds.

In 1867 Tate read a paper before the Geological Society of London
' On some Secondary Fossils from South Africa,' '■' in which he
described several species of plants collected by Drs. Eubidge and
Atherstone from rocks to which the author of the paper applied the
ierm Uitenhage series. The majority of Tate's type-specimens are
now in the Museum of the Geological Society at Burlington House,
and I take this opportunity of expressing my thanks to the Council
for allowing me to borrow several of Tate's plants for purposes of
examination and redrawing. Four of Tate's type-specimens are
represented in pi. v., figs. 1, 3, 4, 5.

The Uitenhage flora, as descril)ed in the paper of 1867, comprises
the species enumerated in the following list. My examination of the
original specimens, and a comparison of them with the more abun-
dant, and in many cases more satisfactory, material recently collected
have led me to make several alterations in Tate's nomenclature ; the
names used by Tate are given in the left-hand column and opposite
each of his names I have added in the right-hand column the
designations adopted in the present paper.

Tate (1867).

Palaiozamia [Otozamites) recta Tate.
Tate's pi. v., figs. 7a, Ih.

Palaozamia (Podoza)iute^) inorrisii Tate.
PI. v., fig. 4.

PaUrozamia rnhuhici Tate.
PL v., figs, -da, -61,.

Palcsozainia {\e\ Pteropliyllum) africaiia

Tate.
PI. v., fig. 5.

Seward (1903).

Z<iinites recta (Tate).t
PL iii. ; pi. vi., figs. 8-12.

Zamites morridi (Tate).

PI. v., fig. 4. (Type-specimen of Tate,

Museum of Geological Society, No.

11,108.)

Zatnitci^ ruhhlijci (Tate).
PI. v., fig. 3. (Type-specimen of Tate,
Mus. Geol. Soc, No. 11,109.)

Zamites africuna (Tate).
PI. v., fig. 5. (Type-specimen of Tate,
Mus. Geol. Soc, No. 11,110.)

• Tate (67).

f In cases where the generic designation has been altered I have followed the-
convenient rule of enclosing the name of the author of the species in brackets.

/^/,

' M./

Fossil Floras of Cape Colony.

Tate (1867).

Pecopteris atherstonei Tate.
PI. V. , fig. 2a, 2b.

Pecopteris ruhidgei Tate.
PI. v., figs, la, ih.

Pecopterh africana Tate.
PI. vi., figs, la, 1/;.

Asplenitea lobata Old.

Sphenopteris antipochim Tate.
PI. vi., fig. 3.

Cyclopterh jenk'ui^iana Tate,
pi. vi., fig. 4.

Seward (1903).

Cladophlehis denticulata (Brongii.) forma

atherstonei.
PI. vi., figs. 16, 17.

Cladophlehis denticnlaUi (Biongn.) forma
atherstonei.

Cladophlehis denticulata (Brongn.) forma
atherstonei.

Cladophlehis hroivniana (Dunk.)
PI. ii., figs. 1-4, 6.

Onychiopsis mantelli (Brongn.).
PI. i. ; pi. v., fig. 1.

Tate's type-specimen refigured, pi. v.,
fig. 1. (Mus. Geol. Soc.No. 11,114. >

Cycadolepis joikinsiana (Tate).
PL iv., figs. 3-6. Text-figm-e 2.

"Portions of a Coniferous stem closely ' Tlrachyjthylliim sp.

PI. vi., figs. 13, 18.

allied to Athrotaxites indicus Old."

" Parts of stems . . . probably Cycada-
ceous."

" Ovules of Palceozamid.^^

" The under surface of the base of a
cone, perhaps the same " [Palmo-
zamia'].

Benstedtia sp.

PL v., fig. 2. Text-figure .5.

Carpolithcs sp.

Araucarites rogersi sp. nov.
PL vi., figs. 4-7.

The Uitenhage series is described by Tate as Jurassic in age ; he
compares the Cycads with species from the Inferior Oolite plant-
beds of the Yorkshire coast, and considers two of the Pecopteris
species as closely allied to P. indica from the Upper Gondwana beds
of the Rajmahal Hills. In 1856 Sharpe '•' described several species
of Mollusca from the Secondary rocks of the Sunday and Zwartkop
Rivers collected by Dr. Atherstone and Mr. A. G. Bain, which he
found to resemble most nearly species from the middle and lower
part of the Oolitic series. The Jurassic age of the Uitenhage series
is accepted by Stow f in his paper of 1871, and H. F. Blanford \ in
1875 referred to the flora, on the authority of Tate, as resembling
that of the Eajmahal beds of India. The evidence as to geological

* Sharpe (56), p. 202.

t Stow (71).

+ Blanford (75), p. 534.

4 Annals of the South African Museum.

liorizon furnished by species of Tr'ujonice. is regarded by Lycett ■-' as
pointing to a Cretaceous age.

In 1882, as the result of an examination of a collection of shells,
Holub and Neumayr i classed the Uitenhage series as Neocomian,
thus agreeing with Lycett rather than with the earlier authors.
Professor Eupert Jones,]: in an article on the Geology of South
Africa, published in Nature, adopts Tate's opinion as to the Jurassic
age of the plant-beds. The verdict of Holub and Neumayr is
accepted by de Lapparent,§ who classes the Uitenhage series with a
portion of the Potomac formation of the Eastern United States as
Neocomian. In Oldham's edition of the 'Geology of India,' || by
Medlicott and Blanford, the flora of the Uitenhage series is
described as " distinctly related to that of the Eajmahal Group,
though the resemblances are not sufficient to establish a contem-
poraneity of origin." Messrs. Eogers and Schwarz, after discussing
the statements of previous writers, conclude that " the only course
at present open to us is to regard the age of the Uitenhage series as
Upper Jurassic." *\

The evidence afforded by the collection of plants described in the
following pages — a collection rich in specimens but comparatively
poor in species — points to a Wealden rather than to a Jurassic age :
but this question is dealt with at greater length in the sequel.

II.— DESCEIPTION OF SPECIMENS.

Group FILICALES.
Family POLYPODIACE^.

Genus ONYCHIOPSIS Yokoyama.

This genus was instituted by Yokoyama,** a Japanese author, for
a species originally described by Geyler f f from the valley of Tetori-
gawa in the Province of Kaga, Japan, as Thyrsoptcris elongata.

* Lycett (79), p. 229.

t Holub and Neumayr (82). See also Neumayr (87), ii., p. 330.

+ Jones (84), p. 554. § de Lapparent (00), p. 1267.

II Medlicott and Blanford (93), p. 205.

^\ Rogers and Schwarz (01), p. 17.

•* Yokoyama (89), p. 26. ft Geyler (77), p. 224.

Fossil Floras of Cape Colony. 5

The use of the name Thyrsopteris as a generic designation for fossil
ferns has unfortunately been adopted by several authors solely on
the ground of a resemblance between the fragments of sterile fronds
with those of the single existing species, Thyrsopteris elegans, a type
confined to the island of Juan Fernandez. Yokoyama's discovery of
fertile pinnae of Geyler's species exhibiting a striking resemblance to
those of the recent genera Onychium and Gymnogramme, led him to
suggest the name Onychiopsis as a convenient designation for ferns
possessing the Onychium type of fertile frond.

Onychiopsis mantelli Brongn.

Plate I.

Plate v., fig. 1.

1824. Hymenopteris psilotoides Stokes and Webb, Trans. Geol.

Soc, [ii.] vol. i., p. 423, pi. xlvi., fig. 7.
1828. Sphenopteris Mantelli Brongniart, Prodrome, p. 50.
1828. Sphenopteris Mantelli Brongniart, Hist. v6g. foss., p. 170,

pi. xlv., figs. 3-7.
1833. Sphenopteris Mantelli Mantell, Geol. S.E. England, p. 241.
1836. Cheilanthites Mantelli Goppert, Foss, Farrnkrt., p. 231 ;

Nova Acta Ac. Caes. Leop.-Car., vol. xvii. Supp.
1839. Cheilanthites denticulatus Eoemer, Verstein. Ool.-Geb., p. 9,

pi. xvii., fig. 1.
1846. Sphenopteris Mantelli Dunker, Wealdenbildung, p. 2, pi. i.,

fig. 4a.
1846. Sphenopteris Boemeri Dunker, loc. cit., p. 3, pi. i., figs. 3, 4, 5.
1846. Sphenopteris tcncra Dunker, loc. cit., pi. viii., fig. 5.
1846. Confervites fissiis Dunker, loc. cit., pi. i., fig. 1.
1852. Sphenopteris Mantelli Ettingshausen, Abb. k.-k. geol. Keichs.,

p. 14, pi. iv., figs. 3, 4.
1867. Sphenopteris antipodum Tate, Quart. Journ. Geol. Soc,

vol. xxiii., p. 139.
1869. Sphenopteris {Davallioides) Mantelli Schimper, Trait, pal.

veg., vol. i., p. 393.
1871. SpJienopteris Mantelli Schenk, Palaeontographica, vol. xix.,

p. 208, pi. xxiii., figs. 1-8.
1871. Sphenolepis Kurriana Schenk, pi, xxv., fig. 6, loc. cit., p. 243,

pi. xxxviii., fig. 2.
1875. Sphenopteris Mantelli Schenk, Palaeontographica, vol. xxxiii.,.

p. 158, pi. xxviii., fig. 12.

6 Annals of the South African Mitscuvi.

1877. Tliyrsoptcris dongata Geyler, Palaeontographica, pi. xxx.,

fig. 5, pi. xxxi., figs. 4, 5.

1878. Sphcnopteris Roemeri Dupont, Bull. Ac. E. Belg., vol. xlvi.,

[ii.] p. 396.
1881. S'phenopteris 2Iantclli Heer, Secc. Trab. Geol. Portugal,

1881, p. 12, pi. xi., figs. 1-5 ; pi. xii., figs. 2h and 2hh.
1881. Sphenopteris valdensis Heer, loc. cit., p. 14, pi. xv., figs. 9,

10, and 13; pi. xvi., fig. 6b. (The other figures are

doubtful.)
1889. Sphenoptcris Mantclli Fontaine, Potomac Flora, p. 91, pi. i.,

figs. 1 and 2.

1889. Onychiopsis elongata Yokoyama, Journ. Coll. Sci. Imp,

Univ. Japan, vol. iii., pi. i., p. 27; pi. ii., figs. 1-3;
pi. iii., fig. 6d ; pi. xii,, figs, 9, 10.

1890. Onychiopsis elongata Nathorst, Denksch. k. Ak. Wiss, Wien,

pi. ii., fig. 6 ; pi. v., fig. 3 ; pi. vi., fig. 5.
1894. Onychiopsis elongata Yokoyama, Journ. Coll. Sci. Univ.
Japan, vol. vii., pt. iii., p. 215 ; pi. xx., fig. 8 ; pi. xx.,
figs. 1, 4.

1894. Sphenoptcris [Davallia) Mantelli Saporta, Flor. Portugal,

pi. xiii., figs. 5, 6 ; pi. xxiii., figs. 1, 2 ; pi. xxviii., fig. 2 ;
pi. xxix., fig. 1 ; pi. xxx., figs. 9, 10 ; pi. xxxi., figs. 1, 2.

1895. Sphenoptcris Mantelli Ward, Potomac Form. p. 392.

1896. Onychiopsis Mantelli Ward, Lower Cret. Europe and

America, p. 483.
1896. Onychiopsis elongata Ward, ibid.

The following definition of Onychiopsis mantelli is quoted from the
first part of the Catalogue of Wealden Plants in the British Museum. ='=
The fertile segments represented in the EngHsh specimens have not
so far been found in the South African material. " Frond tripinnate,
ovate lanceolate, rachis winged and prominent ; pinnae lanceolate,
alternate, approximate, given off from the main rachis at an acute
angle. Pinnules alternate, narrow, lanceolate acuminate, uninerved,
of nervation type CocJiopteridis ; the larger ones serrate and gradu-
ally passing into pinnae, with narrow ultimate segments."

[" Fructification in the form of sessile or shortly stalked linear
ovate segments with rugose surfaces, and terminating usually in a
very short awn-like apical prolongation."]

The name Siihcnopteris mantclli was substituted by Brongniart I

* Seward (94), p. 4:}. f Brongniurt (28), p. 50 ; (28^') p. 170.

Fossil Floras of Cape Colony. 7

for Hymenopteris pisilotoides, formerly used l)y Stokes and Webb *
for fragments of a fern from the English Wealden. In 1894 the
discovery of several fertile pinnae demonstrated the close relationship
of the species to the Japanese fern described by Yokoyama as
Onycliiopsis cloiigata (Gey.). The material from South Africa is
unfortunately represented solely by sterile fronds, and we are there-
fore unable to found a complete specific diagnosis on the Uitenhage
specimens. It is as a rule inadvisable to employ a generic name
implying relationship with existing types unless the comparison is
supported by the evidence of fructification characters. It must be
admitted that so long as we have only sterile examples before us
there necessarily remains an element of doubt as to the correct-
ness of our interpretation of the specimens. In the present case
there is so close a resemblance as regards the habit of the fronds
and in the form of the ultimate segments between the Uitenhage
plant and such European species as Onycliiopsis mantclli (Brongn.),t
0. capsidifera(Ve\.),\ andthe Japanese type, 0. elongata (Gey.), ^, that
I have ventured to adopt Yokoyama's generic name Onycliiopsis.

The fern figured by Tate as Splienopteris antipoclum has been
referred to in a previous work as apparently identical with the
European specimens of Onycliiopsis mantelli.\\ The drawing of
'Tate's type-specimen, "i reproduced in pi. v., fig. 1, shows rather
more clearly than the figure published in 1867 the identity of
Splienopteris antipodum, at least as regards sterile pinnae, with
Onycliiopsis manteUi. In Part I. of the Catalogue of "Wealden
plants in the British Museum I figured a fragment of a frond as
Onycliiopsis elongata, retaining Geyler's specific name as charac-
terising a type of frond with pinnules of slightly greater breadth
than those of Onycliiopsis mantelli ; the separation of the two
species was, however, made with some hesitation and with the
admission that both forms might eventually have to be included
under one term."* Since the publication of the Catalogue a much
larger and more complete specimen of Onycliiopsis has been obtained
from the Wealden beds of Sussex with pinnules of slightly broader
type, which confirms my suspicion that the recognition of two
■distinct species was based on insufficient evidence.

The plant figured by Geyler in 1877 as Tliyrsopteris elongata, and

* Stokes and Webb (24), p. 423. f Seward (94), p. 41. pi. ii.

I Velenovsky (88), pi. i., figs. 6-12.

§ Geyler (77), pi. xxx., xxxi. :. Seward (94), p. 44.

•f Museum of the Geological Society of London, No. 11,114.

** Seward (94), p. .5-5, pi. ii., fig. 2.

S Annals of the South African Musenm.

afterwards described and illustrated more fully by Yokoyama as'.
Onycliiopsis clongata, is spoken of by the latter author as the chief
characteristic fossil of the Japanese Flora of Kaga, Hida, and
Echizen, which is referred to the Bathonian stage of the Inferior
Oolite. Fragments of the same type are figm-ed by Yokoyama,
also from the plant-beds of Kozuke, Kil, Awa, and Tosa referred
to the Neocomian series.* The Japanese type differs from the
European Omjchiojjsis mantelli in its slightly broader segments,,
but the difference is hardly enough to warrant a specific separation.
The fact that both forms possessed the same type of fertile pinnae
demonstrates their close relationship, and such small differences
as are exhibited by the sterile fronds are not greater than analogous
variations met with in species of existing ferns.

The specimens figured by Nathorst f from the province of Tosa, in
Japan, are undoubtedly specifically identical with Geyler's species,
and I believe also with the Uitenhage plant.

Eeference is made in the synonomy to the figures of Onychiopsis-
mantelli, published by Schenk and other authors from European
Wealden rocks ; it is unnecessary therefore to do more than draw
attention to the abundance of this well-marked type in Wealden
floras. The same species is recorded also by Fontaine I and Ward §
from the Potomac formation of North America ; some of the ferns^
refei'red ))y the former author to Tlii/rsopteris, without a fragment'
of a fertile pinna to justify the assumed relationship with the recent
genus of Cyatheaceaj, are in all probability identical with Onycliiopsis
viantelli. We may quote Tkyrsopteris rarinervis, T. insignis, T..
micropliylla as examples of Fontaine's species which bear a close
resemblance to Onychiopsis mantelli.\\ A fern described by Heer-
as Asplenium dicksonianum *'. from the Kome beds of Greenland, if
not identical with the Uitenhage species, is probably a closely allied
type. In the late Marquis of Saporta's monograph on the Mesozoic
Floras of Portugal,** Onychiopsis viantelli is recorded from several
localities as Sphcnoptcris (Davallia) mantelli, but the evidence on
which a relationship to Davallia is founded seems to me inadequate ;
Saporta speaks of the species as a characteristic Wealden fern, but
fragments found in somewhat older beds point to the existence of

* Yokoyama (94), p. 'ilo, pi. xx., xxi.
t Nathor-st (90), pi. ii., fif,'. (i ; pi. v., fig. 3; pi. vi., fig. 5.
+ Fontaine (89), pi. i. § Waixl (95), p. 392; ('.lO), p. 483.

II Fontaine (89), pi. xlii., xliii., xliv., xlv.
If Heer (75'), pi. i., figs. 1-5.
** Saporta (94), pi. xiii., xxiii., xxviii.-xxxi.

Fossil Floras of Cape Colony. 9>

the same type before the close of the Jurassic period. It is note-
worthy, however, that the specimens from the Jurassic locaUties are
very small, and several of them can hardly be referred with con-
fidence to Onycliiopsis mantelli, which is represented by many
undoubted examples from the Wealden of Portugal as demonstrated
by the figures of Heer '■'■ and Saporta.

Plate I., fig. 1 (428c).
Portion of a large frond, approximately 17 cm. in length, with long
spreading pinnae. The slender rachis shows narrow lateral wings,,
and in places a prominent woody axis ; the pinnae are decurrent on
the rachis, and the upper branches of the frond are given off at a
smaller angle than the lower. Longest pinnae 8 cm. long, tapering
gradually to a serrate acuminate apex. Veins obscure. Specimen
434a is the reverse of 428c.

Plate I., fig. 2 (391c), l^ nat. size.

A single pinna bearing in the lower part linear serrate pinnules,
and in the apical region shorter entire segments. The specimen
from which the drawing was made differs from the frond repre-
sented in fig. 1 in its slightly narrower and more delicate pinnules,
and is identical with the typical European examples of the species.

Plate I., fig. 3 (407c), x 2.
The complete specimen of which a small portion is represented in
the figure, twice natural size, is 8 cm. in length, and agrees in all
respects with the English and German form of Omjchiopsis mantelli.
The single pinna shows the entire and shorter segments charac-
teristic of the apical region of a frond.

Plate L, fig. 4 (401c), x 2.
A portion of the rachis and the base of two pinnse illustrating the
broader type of lobed pinnules. This fragment shows very clearly
the winged character of the rachis and the prominent woody axis.

Plate v., fig. 1. (Type-specimen of Tate's S2:)heiiopteris antipodum ;
Museum of Geological Society of London, No. 11,114.)

This specimen from Geelhoutboom is, I believe, clearly identical
with Onycliiopsis viantclU (Brongn.), at least as regards the vegeta-
tive characters of the frond. The rachis is traversed by a narrow

* Heer (81), pi. xi., xii.

10 Annals of the South African Museum.

woody axis with lateral wings, and gives off spreading pinnge bearing
slender linear segments the lamina of which is divided into acuminate
teeth,

Ot]ier Specimens : 433c. — This specimen shows very clearly the
narrow pinnaj with entire segments (c/. fig. 3) in the apical region of
the frond; also the gradual transition from the narrow oval segments
to larger segments with a serrate mai'gin. 140c, 320c (on the reverse
side is a good impression of a Cycadolepis jenkinsiana), 389c, 390c
{with seeds of Araucarites rogersi), 392c, 411c, 429c, 430c.

Genus CLADOPHLEBIS Brongniart.

(Of doubtful Family-position.)

Brongniart proposed this generic name in 1849 " for certain forms
of Pecopteris, previously included in the group Neuropterides. The
genus is founded on sterile fronds only, and is applied to ferns which
•cannot be referred to a particular family ; the name Cladophlebis is
usually made use of for Mesozoic species which agree in habit with
Palaeozoic representatives of Pecoptcris. Numerous fossil ferns have
been described by authors as species of Aspidium, which, in the
absence of sori or sporangia, have no claim to be designated by the
name of a recent genus.

Cladophlebis browniana (Dunk.).
Plate II., figs. 1-4, 6.

1846. Pecoptcris Browniana Dunker, Wealdenbildung, p. 5, pi. viii.,

fig. 7.
1849. Pecoptcris Broioniana Brongniart, Tableau, p. 107.
1852. Alcthoptcris Bcichiana Ettingshausen, Abh. k.-k. geol. Eeichs.,

vol. i., Abth. iii., No. 2, p. 17.
1871. Pecoptcris Broivniana Schenk, Palaeontographica, vol. xix.,

p. 215, pi. xxvi., figs. 2, 2a.
1877. Pecoptcris exiliformis Geyler, Palseontographica, vol. xxiv.,

p. 226, pi. xxx., fig. 1.
1889. ? Pecoptcris Browniana Fontaine, Potomac Flora, pi. xxiii.,

figs. 2-7.

* Brongniart (49), p. 25.

Fossil Floras of Cape Colo)uj. 11

1890. Pecopteris Gcylcriana Nathorst, Denkschr. k. Ak. Wiss. math.-

nat. CI., vol. Ivii., p. 48, pi. iv., figs. 2-6.
1890. Pecopteris cf. Broioniana, ibid., pi. v., tig. 5.

Frond bipinnate, passing in the lower part of the frond to tri-
pinnate. Eachis comparatively broad and strong, from which linear
pinnae, tapering to an acuminate apex, are given off at a wide angle.
Pinnules short and broad, attached by the whole broad base to the
axis of the pinna ; the lower margin strongly arched, the upper face
almost straight or slightly concave. A well-marked midrib from
which spring a small number of secondary veins at a wide angle.
The pinnules are for the most part entire, passing in the lower
pinnas into lobed or even pinnate segments.

The Uitenhage specimens include numerous examples of bi-
pinnate and tripinnate fronds which were referred by Tate in his
paper of 1867 * to a species previously figured by Oldham | from the
Eajmahal Hills of Bengal as Pecopteris {?) lobata. A comparison of
the fragments represented in pi. ii., figs. 1-4 and 6, with the
specimen figured by Oldham reveals a very close resemblance,
which, as Tate suggested, may amount to specific identity. On the
other hand, there is, I think, an even closer agreement between the
Uitenhage fern and that described from European Wealden strata
as Cladoplilehis dunkeri (Schimp),]: a form originally nained by
Dunker in 1846 Pecopteris polymorpha.l The fern named by
Dunker Pecopteris broiuniana \\ differs in no essential features from
Cladophlcbis dunkeri, and a re-examination of specimens from the
Wealden beds of the Sussex coast leads me to give up my previously
expressed opinion that both specific names should be retained. The
difference between the fragment represented in pi. vii., fig. 4, of
my ' Wealden Flora,' Part I., and that shown in pi. vii., fig, 3,1i is,
I think, not more than may be found on a single frond of the same
species. Seeing that the name brotoniana was established before
the specific name dunkeri, it should be employed in preference to
Schimper's designation.

The accurate separation of sterile fronds of the type represented
in pi. ii., figs. 1-4, 6, is a hopeless task; among recent ferns this
form of leaf recurs in several genera, and in the absence of fertile

* Tate (67), p. 146.

t Oldham and Morris (63), pi. xxviii.-xxx. ; Feistmantel (77), pi. xxxvi., fig. 3.

J Seward (94), p. 101, pi. vii., fig. 3. § Dunker (46), pi. vii., fig. 5.

II Ibid., pi. viii., fig. 7. 1 Seward (1)4).

12 Annals of the South African Museum.

examples it might well be impossible to arrive at a generic much
less specific determination.

So far as vegetative characters are concerned, the fern described
by Yokoyama from Japan as Pecopteris hrowniana " appears to be
identical with the South African plant ; a frond of very similar if not
identical form is figured by the same author from a somewhat lower
horizon as Pecopteris exilis Phill.,i an Inferior Oolite species since--
placed in the genus Kluhia.\ Nathorst has also figured portions of
fronds from Japan which he regards as probably identical with his-
species Pecopteris geyleriana. It is, however, not improbable that
the specimens represented in his pi. iv., figs. 2, 6, are specifically
distinct from the type-specimen of P. geyleriana ;% be that as it.
may, I am unable to distinguish the two examples shown in
Nathorst's figs. 2 and 5 from the Uitenhage fern. In addition to
the records of ferns believed to be identical with Cladophlehis^
broivniana mentioned in the above synonomy, there are others
which may be referred to as presenting a close resemblance possibly
amounting in some cases to specific identity.

Eeference has already been made to Oldham's Indian species
Pecopteris lobata, as a fern with the same form of frond as that of
Claclophlehis broivniana; the specimen figured by Feistmantel as-
Dicksonia [Splienopteris) bindrabunensis \\ may perhaps represent
the fertile form of Pecopteris lobata. Dicksonia coriacea, a Chinese^
species recorded by Schenk affords another example of the C.
Broiuniana form of frond."' Some of the fragments figured by
Saporta='=* from Neo-Jurassic rocks of Portugal maybe identical with
the African plant, and a comparison may also be made with
Aspidium hctcropliyllum and Clculophlebis distansW described by
Fontaine from the Potomac beds. Aspidium montanense figured by
the same author \ l from the Great Palls of Montana represents-
another very similar species.

Another fern of precisely similar habit to Gladophlebis broivniana
and hardly distinguishable from it is a species recently placed in
the genus Coniopteris, on the evidence of fertile specimens, but
originally described as Ncuroptcris anjuta L. and H.§§ This Inferior

* Yokoyama (94), p. 218, pi. xxiv., figs. 2, 3; pi. xxvii., figs. 1-5.

t Ibid. (89), pi. i., figs. 8-10. + Raciborski (91).

S Nathorst (90). || Feistmantel (77), pi. xxxvii., fig. 2.

•i Schenk (88), pi. lii., figs. 5, 6.

*• Saporta (94), pi. vi., fig. 1 ; pi. vii., fig. 5; pi. xi., xii.

tt Fontaine (89), pi. xiii., figs. 4, 5 ; pi. xv., figs. 1-5.

J{ Ibid. (92), pi. Ixxxii., Ixxxiii., and Ixxxiv.

§§ Sewai-d (00^), p. 115, pi. xvi., xvii. ; Lindley and Hutton (34), pi. cv.

Fossil Floras of Cape Colony. 13

•Oolite species probably belongs to the Cyatheaceae, but in view of
the frequent occurrence of this type of frond in different genera of
recent ferns it is unsafe to assume that the South African type
possessed sori and fertile segments like those of the Inferior Oolite
•species.

Plate II., figs. 1, la (343c).

The apical portion of a frond bearing linear and slightly falcate
pinnte with short and broad pinnules and entire apices. The
venation is more accurately shown in fig. la ; the enlarged draw-
ing also illustrates the incomplete separation of the pinnules as
•contrasted with their more complete separation in the lower pinnoe.

Plate II., figs. 2, %i (344c).

Similar to fig. 1, but showing more clearly the comparatively stout
rachis ; the pinnules represented in fig. %i differ from those shown
in fig. la in their more complete separation and in their greater
length. The venation is clearly seen in the enlarged pinnules, also
the shorter and almost deltoid form of the basal pinnule on the
lower side of the pinna.

Plate II., fig. 3 (364c).

The specimen, of which a small piece is represented in the figure,
consists of fairly large portions of fronds similar to that drawn in
fig. 4. Fig. 3 illustrates the attachment of a pinna to the rachis,
also the longer type of pinnule — 4-5 mm. — characteristic of the
lower branches of a frond.

Plate II., figs. 4, 4fl, 4/j (342c).

The rachis has a length of 7 cm., and the longest pinna
4-5 cm. This specimen shows very clearly the characteristic habit
of the frond ; the stout rachis giving off opposite or alternate pinnae
of linear acuminate form with falcate segments, most of which have
entire margins (fig. 4a), while others (fig. 46) in the lower part of
the leaf show signs of an incipient subdivision of the lamina into
rounded lobes — a condition exhibited in a more pronounced form
in fig. 6.

On the same piece of rock is an impression of a piece of
Benstedtia, and on the reverse side a portion of the rachis of
Zamites recta shown in pi. iii., fig. 1.

14 Annals of the South African Museum.

Plate II., tig. 6 (361c).

A portion of a pinna, slightly enlarged, from a lower part of a
frond than the pinnule outlined in fig. ^h.

Other Specimens : 339c, 340c, 349c, 350c, 351c, 358c, 359c, 363c,
366c, 369c, 373c.

Cladophlebis denticulata (Brongn.), forma athekstonei.

Plate VI., figs. 16, 17.

1867. Pecoptcris Athcrstonei Tate, Quart. Journ. Geol. Soc, vol. xxiii.,,

p. 145, pi. v., figs. 2a, 26.
1867. P. liiihidgei, ibid., p. 146, pi. v., figs, la, lb.
1867. P. africana, ibid., p. 146, pi. vi., figs, la, lb.

Frond bipinnate, rachis broad, giving off pinnae at a wide angle ;
the branches or pinnse bear crowded linear pinnules attached by the
whole of the broad base to the comparatively broad axis ; the linear
pinnules reach a length of slightly over 3 cm., apices bluntly
pointed ; midrib strong, giving off secondary veins at a wide angle,
which branch dichotomously as they pass obliquely upwards to the
edge of the lamina. Fertile pinnules unknown in the South African
form.

The most abundant plant in the material collected two miles east
of Herbertsdale is a fern characterised by linear pinnules of the
Cladophlebis type, which, owing in part to the friable nature of the
rock, is seldom preserved except in small broken pieces of pinnae.
Fragments of Tceniopteris fronds (pi. ii., fig. 5) are constantly found
in association with this form of Cladophlebis. The piece represented
in fig. 16, pi. vi., shows the venation very clearly preserved, and
with fig. 17 illustrates the form of the pinnules. There can be no
doubt that this fern is the one named by Tate Pecopteris athcrstonei ;
an examination of Tate's type-specimens leads me to regard the
smaller form named by this author Pecopteris rubidgei as a frag-
ment of the same species. In all probability we may safely include
Tate's name P. africana as another synonym under P. athcrstonei.
No trace of sori or fertile pinnae has been found, and we must thex-e-
fore be content to leave the plant as a member of the Filices of
uncertain family position. As regards the question of nomenclature :
there is no more common type of fern frond among recent genera
than that of the bipinnate leaf with hncar pinnules, attached by the

Fossil Floras of Gape Coloiuj. 15

whole of the base to the pinna-axis, like those represented in figs. 16
and 17, pi. vi. Among existing species we find it in Onoclea
germanica W., Alsophila jwujluthiana Mett., Cyathca dealbata Sw.,.
Gleichenia glauca Sw., Sadler ia cyatheoidcs Kaulf., Pteris arguta
Ait., Todca harhara Moore, also in species of Folypodium, Asplenium,
and other ferns. In previous publications dealing with Mesozoic
floras I have drawn attention to the wide geographical range of a
similar type of frond during the Ehaetic, Inferior Oolite, and
Wealden periods.''' The form named by Brongniart Pecopteris
dcnticulata and now placed in the genus Cladophlehls occurs as an
element of Jurassic floras in all parts of the world. Fertile pinnae
are rarely found, and none that have come under my observation
show recognisable sporangia, but attention has elsewhere been
drawn to the close resemblance — as regards the form of the sori
and their distribution on the lamina— between some English
specimens and fertile pinnae of Todea barbaraA This resemblance
is, however, insufficient to serve as a safe guide to affinity, and all
that we can say is that it is not improbable that some at least of the
fronds of the type represented by Cladopldchis dcnticulata may be
members of the OsmundaceiE. In a previous work, l^efore the
^-•possible relationship to the Osmundacese occurred to me, I included
Gladophlchis dcnticulata as a doubtful member of the Polypodiacege.
A fertile specimen of Todca australis (Morr.) from Australia was
found sufficiently well preserved to enable Kenault j to examine the
sporangia, which he describes as agreeing closely with those of
recent Osmundaceee.

It is obviously impossible to discriminate between the numerous
sterile fragments of this form of frond from various geological
horizons, ranging from Ehaetic to Lower Cretaceous, and furnished
by rocks extending from Greenland to Australia and New Zealand,
at least so far as concerns the recognition of distinctive characters
that could reasonably be considered of specific rank. A possible
method to adopt is to use the term Gladophlchis dcnticulata (Brongn.)
as a comprehensive title for ferns, for the most part of Jurassic age,
characterised by bipinnate fronds with linear ultimate segments of
the shape arid with the venation characters shown in figs. 16, 17,
pi. vi. We may in particular cases employ a second name, as
denoting a possible variety or form and as an index of locality.
In the present instance we may therefore speak of the Herbertsdale

* Seward (94), p. 95; (00), p. 18; (00^), p. 134.

t Seward and Ford (03). I Eenault (83), p. 81, pi. xi.

16

Annals of the South African Museum.

fern as Cladophlehis dcnticulata forma atherstonei. The epithet
denticulata may in some instances prove misleading, but the presence
■or absence of fine denticulations is in itself hardly a character of
primary taxonomic value. Moreover, in imperfectly preserved
specimens so small a point is not easy to determine. The figures
of English East Yorkshire specimens * of Brongniart's type bear a
very close resemblance to the fragments of the South African plant ;
similarly Pecopteris indica 0. and M.f from the Eajmahal series of
Bengal, Asjjlenium argutuhim Heer, from China,:]: Fontaine's Clado-
phlehis oblongifolia from the Potomac formation, § Pteris frigida and
P. longipennis Heer,]] and Cladophlehis steioartiana Hartz *! are some
of the many examples of fern fronds that might well be grouped
imder Cladophlebis denticulata.

The following list serves to illustrate the geological and geographical
range of ferns comparable with, and no doubt in part identical with,
Brongniart's Inferior Oolite type. The references to authors are
given in full in the Bibliography.

Names employed by
Authors quoted.

Cla(lo2>hlebis denticulata. .

Pecojitcris indica

Cladophlebis steicurtiania

Pteris frigida

P. alhertini

CladojMehis oblongifolia. .

Asplenium distans

Asplenitnn argutulum ....

Alethoj)teris australis

Polijpodium hochstetteri ..

Asph-iiiuni distans

Cladophlebis dcnticulata . .

Geographical range.

England (E. Yorks)
India (Bengal)
Greenland

Greenland

Bohemia

Eastern N. America

Canada (Eocky

Mountains)
China
Australia
New Zealand
Japan
Poland

Approximate
geological horizon.

Author.

.Jurassic
Jurassic
Jurassic

Lower Cretaceous
Lower Cretaceous
Wealden — Upper

Jurassic
Lower Cretaceous

Jurassic
Jurassic
Wealden ?
Jurassic
Jurassic

Seward (00) (00^)

Oldham & Morris (63)

Feistmantel (77)

Hartz (96)

Heer (82)

Fric and Baver (01)

Fontaine (89)

Dawson (85)

Schenk (83)
Feistmantel (90)
Unger (G4)
Yokoyama (89)
Raciborski (94)

Plate VI., figs. 16, 17 (170d).

The specimen shown in fig. 16 is one of several fragments of
pinnae bearing linear pinnules approximately 3 cm. in length. The
distal portion of the pinnule represented in fig. 17 illustrates the
form of the apex which is not seen in fig. 16. The Herbertsdale

• Seward (00), pi. xiv., xv., xx.
\ Schenk (83), pi. xlvi.-xlviii., lii,
II Heer (82), pi. x.-xiii.

t Oldham and Morris (62), pi. xxvii.
§ Fontaine (89), pi. vii., figs. 8-5.
^\ Hartz (90), pi. xi.

Fossil Floras of Cape Colonij. 17

rock, ill wliich this fevn is very abundant, contains also numerous
fragments of Tamiopteris fronds and pieces of stems or probably
broad rachises.

Otlicr Specimens : 173d, several imperfect portions of pinnsB and
pinnules, also a piece of what is probably a broad rachis or petiole :
with fragments of Tcenioptaris. 178d, 179d, 18od, 357c, smaller pin-
nules slightly larger than 1 cm., 368c smaller pinnules with clearly
marked venation, A.

Genus SPHENOPTEKIS Brongniart.

Sphenoptekis fittoni Sew.

Plate II., figs. 7, 8.

1836. Sphoioptcyls gracilis Fitton, Trans. Geol. Soc, vol. iv., ser. ii.,

pt. ii., p. 103.
1849. Pacliijpteris gracilis Brongniart, Tableau, p. 107.
1852. Pacliijpteris gracilis Ettingshausen, Abb. k.-k. Geol. Reichs.,

vol. i., Abth. iii., No. 2, p. 21.
1864. Asplenium palcBopteris linger, Reise Fregatte Novara, vol. i.,

Abth. ii., p. 3, pi. i., fig. 4.
1893. ? Spheiioptcris delgadoi Saporta, Eev. Gen. Bot., vol. v., 1893,

p. 270, pi. iv., fig. 5,

Frond ovate lanceolate, bipinnate, possibly becoming tripinnate in
the lower part ; pinnae linear acuminate, alternate, springing at a
wide angle from the rachis ; pinnules ovate acuminate with entire
margins passing gradually into more or less deeply Jobed segments
obliquely attached to the axis of the pinna on which their lower
margin is decurrent. Venation of the SpJienoptcridis type.

The specimen on which this species is founded was described in
1827 by Fitton from the English Wealden rocks as Sphenopteris
gracilis/'' but before his paper was published by the Geological
Society of London the specific name gracilis had been applied to a
Carboniferous example of Sphenopteris. Fitton's type-specimen,
now in the Museum of Practical Geology, Jermyn Street, London,
is clearly identical with examples more recently found by the late
Mr. Rufford on the Sussex coast near Hastings, and I have little
doubt that this English fern, so far as characters based on sterile

* Fitton (.36).
0

18 Annals of the South African Museum.

fronds enables me to form an opinion, agrees too closely with the
Uitenhage form to justify specific separation.

The entire form of pinnule, represented in fig. 8, pi. ii., is found in
the upper pinnae of a frond, and every gradation is met with connect-
ing the narrow segments in the apical portion of the fragment repre-
sented in fig. 8 with segments like those of fig. 8a and the more
deeply lobed type shown in figs, la, 7b. In addition to the
species mentioned in the above synonomy, there are various others
with which SphenoiJteris fittoni exhibits a close agreement. Of
these reference may be made to Scleroptcris i)omclii and S. tenuisecta,
figured by Saporta ■'■ from French rocks ; also Sclcroptcris terndsecta
and S. tenclliloha, recorded by the same author from the Jurassic
plant-beds of Portugal. i Scleroptcris vernonensis, figured by Ward .]:
from the Potomac formation, and Fontaine's Sclcroptcris virginica §
are other examples of ferns of similar habit ; also Sphcnoptcris
pilurincrvia, a Portuguese species described by Heer. |1

Plate TL, fig. 7 (249c).

An imperfect and somewhat ol)scure piece of a frond 7"5 cm. in
length, a portion of which is represented in the figure. From a
broad rachis pinnae are given off at almost a right angle. The
habit of the frond is more open than in Cladophlchis hroioniana.
The pinnules of Sphenoptcris fiitoni are distinguishable by their
narrower and more pointed form and more acute lobes. The en-
larged drawings (figs, la, lb), show the form of the pinnules more
clearly ; the venation is obscure, but traces of a midrib can be seen
here and there. On the same piece of rock occur also fragments of
Araucaritcs scales and a good example of Zamitcs rachis.

Plate II., fig. 8 (37Gc).

A portion of a pinna with more clearly preserved pinnules differing
from those shown in fig. 7 in having entire or very slightly lobed
margins (fig. 8a). The venation is clearly seen in this specimen ;
it consists of a central midrib from whicli spring a few secondary
veins at an acute angle.

Other specimens : 347c.

* Saporta (73), pi. xlvii. ; (1)1), pi. lix.-lxi.

t Ihid. (94), pi. ii., iii. \ Ward (95), pi. ii., figs. 1-3.

S Fontaine (89), pi. xxviii., figs. 3, 5. |1 Heer (81), pi. xi., fig. 6.

Fossil Floras of Cape Colony.

19

Sphenopteris sp.
Text-fiffure 1.

The fragment (231c) represented in text-fig. 1 (r) is part of the
terminal portion of a pinna, 2 cm. long, bearing pinnules with well-
defined acutely spi*eading and forked veins of the Sphenojoteris
type (a). It bears a close resemblance to the type of frond illus-
trated by the widely distributed Jurassic species Coniopteris
hymenopliylloides (Brongn.) ; but witliout more evidence it is ad-
visable to leave the fragment as Sphcnoptcris sp.

Other specimens : 234c, 238c.

Genus T.ENIOPTERIS Brongniart.

T.ENioPTEpas sp. [cf. T.EXIOPTEEIS AECTicA (Heer).]

Plate II., figs. 5, 5«.

The fragments of simple Tceiiiopteris fronds are too imperfect to
admit of a satisfactory diagnosis on which to found a species. The
fronds possess a broad midrib from which lateral veins are given off
at right angles or at an angle slightly less than 90° ; these may
pass to the margin as simple veins or bifurcate either close to their
exit from the midrib or in different positions in the lamina. The
leaf is u^sually about 1'5 cm. in breadth and there appear to be
approximately 10 veins per 5 mm.

No specimens have been found showing either the base or apex
of the lamina, and there is no indication of any sporangia. The
abundance of Tceniopteris in Jurassic and Wealden strata and the
close agreement, often amounting to identity, between forms referred
to distinct species renders it inadvisable to apply a distinctive name
to the Uitenhage fragments. A comparison may be made with the

■20 Annals of the South African Museum.

form of leaf named by Heer Olcandra arctica from the Cretaceous
(Kome) beds of Greenland" and with specimens from the Great
Falls of Montana described by Newberryf as " cf. Oleandra arctica
Heer." The type Taniopteris arctica (Heer) is perhaps the most
nearly allied, as regards form and venation, to the South African
leaves. The Eajmahal leaves, described originally by McClelland as
Taniopteris sjyatlmlatal and afterwards by Peistmantel as Angiop-
teridium spatJmlatuin, § represent similar forms. The narrower
leaves from the Eajmahal Hills, figured by Oldham and Morris
and by Feistmantel and I'eferred to McClelland's species, are un-
doubtedly identical with McCoy's TcBniopteris daintreei,\\ abundant in
Jurassic strata of Victoria, and distinct from the Wealden Uitenhage
leaves. Nathorst's species Tceniopteris lundgreni, 1i from Upper
Jurassic beds of Advent Bay, Spitzbergen, is another type with
which the African specimens may be compared.

The European Wealden species, Tceniopteris beyrichii Schk.,"* is
most probably distinct from, although comparable with, such speci-
mens as those shown in pi. ii., figs. 5, 5a.

Plate II., figs. 5, 5a (172d, 193d).

The venation is clearly shown in the figured fragments.
Similar portions of leaves occur abundantly in association with
Cladopldchis denticulata, forma atherstonei, but all are imperfect.
There appear to be about 10 secondary veins per 5 mm. of lamina ;
the average breadth of the leaf is approximately 1 cm.

Other specimens : 168d, 385c.

Group CYCADOPHYTA.

This name has recently been suggested by Nathorst i | for Cycadean
plants which are known only in the form of vegetative organs and
cannot be referred with certainty to either the Bennettitales or
Cycadales — groups characterised by well-marked features exhibited
by their reproductive organs.

* Heer (78), pi. xii., fi-^s. 3-11. f Newberry (91), pi. xiv., fig. 9.

+ McClelland (.50), pi. xvi., fig. 1. J, Feistmantel (77), pi. i.

II McCoy (74), pi. xiv., figs. 1, 2. 1[ Nathorst (97), pi. iii., figs. 1-5.

*♦ Seward (94), p. 152, pi. ix., fig. 3. ft Nathorst (02).

Fossil Floras of Cape Colony. 21

Genus ZAMITES Brongniart.

Zamites kecta (Tate).

Plate III.

Plate VI., figs. 8-12.

1867. Palceozamia {Otozamites) recta, Tate, Quart. Journ. GeoL
Soc, vol. xxiii., pi. v., figs, la, lb.

Fronds pinnate, probably reaching a length of 50-60 cm., rachis
stout and woody.

Pinnae alternate, attached obliquely in two alternating rows by a
slightly narrowed and callous base to the upper face of the frond
axis, in the lower part of a frond they are given off almost at a right
angle, but tow^ards the apex of the pinnate leaf they are inclined at
a much smaller angle ; linear in form, tapered slightly at the swollen
base w^hich w^as probably separated by an abciss-layer from the rachis
as in Encephalartos and certain other recent genera of Cycads ; apex
acuminate and asymmetrical, the upper edge of the pinna almost
straight or slightly falcate, in some cases bending downward
towards the tip, the lower margin curved, bending somewhat
suddenly upward to the apex in the distal third of the pinna. In
shape the pinnte vary from the practically straight form shown in the
upper part of fig. 1 (pi. iii.) to the falcate type, resembling a sword-
bayonet, represented in fig. 3 and in some of the pinnae in fig. 1.
The larger pinnae reach a length of over 6 cm. and a breadth of
9 mm. ; veins numerous and parallel, converging towards the base
of the pinnae where they are frequently forked (fig. la).

The type of frond shown in fig. 1, pi. iii., bears a very close
resemblance to Williamsonia gigas (L. and H.) '■'■' of the Inferior
Oolite of Yorkshire — a species which may w^ell be identical with
Zamites scJimiedelii And.,i and Zamites feneonis Brongn.,| as well
as with such forms as Z. moreaui, Z. clarauallensis, and others
figured by Saporta. j Other similar species are Zamites hohemicus
Vel.,11 from the Lower Cretaceous rocks of Bohemia, Zamites
iburgensis Hos. and von d. Marck.*^ Additional records might be

* Lindley and Hutton (35), pi. clxv. ; Seward (00), pi. v.-viii.

t Andrte (53), \)\. ix., figs. 1-4.

X Saporta (75), pi. Ixxxvii.-xci. ; Heer (7<)), pi. Hi., figs. 2-8.

§ Saporta (75), pi. Ixxxiv.-lxxxv., xciii.

II Fric and Bayer (01), fig. 43, p. 92.

% Hosius and von d. Marck (HO), pi. xliv., fig. 202.

22 Annals of the South African Museum.

quoted of fronds bearing a very close resemblance to Tate's species.
A careful comparison of the African and British examples reveals
a slight diiference in the shape of the pinnas, and this possibly
trivial distinction, taken in conjunction with the fact that no
reproductive shoots of the Williamsonia type have so far been
found in the Uitenhage beds, renders it advisable to use the generic
name Zamitcs in preference to IViUiamsonia and to retain Tate's
specific name. It may be that we shall eventually obtain evidence
of the existence of the Williamsonia shoots in connection with
Zamitcs recta, but until further evidence is forthcoming it is better
to class this species simply as a member of the Cycadophyta than
as a representative of a definite family. The reasons on which the
British species has been assigned to Williamsonia and placed in the
Bennettitales are fully discussed elsewhere."'

Plate III., figs. 1, la (257c, 258c).

A portion of the specimen is shown in the figure, the whole being
14-5 cm. in length. The rachis is partially hidden by the rock
and so appears less than its real size ; the longest pinna is
approximately 5 cm. long. The veins are clearly shown and, as
illustrated in fig. la, they are frequently forked close to their
entrance into the lamina — a feature found also in the segments of
recent cycadean fronds.

Plate III., fig. 2 (288c).

The complete specimen, of which a part only has been drawn, is
9 cm. in length ; it illustrates the narrow form and greater obliquity
of the pinnae in the apical region of a frond.

Plate III., fig. 3 (A).

The single pinna shown in the figure serves as a good example
of the very small angle of attachment of segments in the apical
region of the frond.

A portion of the rachis, in the form of a mould, is seen close to
the base of the segment.

Eachis of Zamites recta (Tate).
Plate VI., figs. 8-12.

The nature of the specimens represented in figs. 8-11 was for
a long time a puzzle, and I am still in doubt as to their significance.
* Sewaid (00), p. 177; (!)7).

Fossil Floras of Cape Colony. 23

At first sight the examples shown in iigs. 8 and 9 suggested surface-
impressions of stems bearing prominent leaf-cushions, but the fact
that no specimens could be found exhibiting more than two rows
of the cushions was a difficulty in the way of regarding them as
fragments of stems. ^Moreover, the two alternating rows were found
in some cases {e.g., fig. 8) to be situated obliquely to one another,
the cushions of one row being almost at right angles to those of the
other series.

It occurred to me that the narrow oval area forming the flat top
of the cushions, and well shown in figs. 8 and 9, agreed in size and
form with the base of the pinnae of Zamites recta (fig. 12, pi. vi.) ;
this led me to make a further search for evidence as to a possible
connection of the rows of cushions with the fronds of the cycad.
The discovery of the specimen represented, rather larger than
natural size, in fig. 11, confirmed this view and clearly demonstrated
that these apparent stems are portions of the rachis of Zamites
recta. The following description of the figured specimens may serve
to elucidate to some extent the nature of these curious fossils.

Plate VI., fig. 8 (289c).

A piece of rachis 4 cm. long, retaining fragments of carbonaceous
matter (not shown in the drawing) on its surface. There are two
alternating rows of prominent cushions obliquely inclined to one
another ; each cushion consists of a flat surface or base exhibiting
a few obliquely longitudinal wrinklings, and this bears a prominent
cushion-like body sloping gradually to a flat top of oval outline
bounded by a fairly prominent rim. The sloping sides of the
cushions are folded into small ribs and bear the impress of numerous
cell-outlines, as shown in the enlarged drawing, fig. 10.

In the lower part of fig. 8 the cushions are absent, and the flat
basal area is more fully exposed ; the transversely elongated
depression a occurs in the interval between adjacent cushions.
Associated with Zamites pinniB, Cycadolcpis, &c.

Plate VI., fig. 9 (328c).
A single row of cushions like those shown in fig. 8.

Plate VI., fig. 10 («).

A small piece of the surface of a cushion enlarged to show the
fine reticulations representing the impressions of cells.

24 Annals of the South African Musewii.

Plate VI., tig. 11 (342c).

A curved piece of rachis slightly enlarged, the actual length being
4-5 cm. In the lower part a portion of a longitudinally striated
woody axis is exposed ; the stout woody character of the rachis is
shown in the outline of a transverse section to the left of the figure.
The two lowest pinnse rest obliquely on the axis. Against the base
of the third pinna the surface of the woody axis is covered with
a thin piece of rock which represents the remains of a cushion ;
a larger and more distinct portion of cushion is seen at c, and at d
the outline of a partially preserved cushion is more distinctly seen.
The position of the pinna to the left of the axis at d shows that the
pinnse abutted on to the side of the cushions. A more clearly
defined cushion is seen at c, but at h the surface-features ai-e
still better preserved. From the upper end of the specimen the
cushions are absent, but the transversely elongated depression is
seen at a.

Plate YL, fig. 12 (286c).

Flattened rachis showing on one side the attachment of a row
of pinnse and on the other the oblique oval scars to which the
pinnae were articulated.

Other Sjicciviois : 255c, portions of two fronds approximately
parallel to one another, with fragments of others A piece of a
rachis or petiole, 7 mm. broad, occurs on this slab of rock which
no doubt belongs to Zamitcs recta. The longest piece of frond is
18 cm. long, and shows that the whole leaf was probably at least
40-50 cm. in length; 259c, 260c, 264c, a piece of a frond about
30 cm. long, with other fi-agments. The small amount of difference
in the position and size of the pinnae at the two ends points to a
frond of considei'able length. On the under surface of the rock
there is a cast of a stem with surface features similar to Benstedtia.
302c, 303c, 345c, 352c. 279c, similar to some examples of Zamitcs
huchianiis (Ett.),''' but differing from that species in the form
of the base of the pinnae. Segments of Araucaritcs scales occur
on the reverse side. 323c, with the impression of a stem (?)
giving off several narrow roots. 334c, with Araucaritcs scales,
346c, 372c.

* Seward {'.)')), pi. iii., iv., viii.

Fossil Floras of Cape Colony. 25

Zamites mokkisii (Tate).

Plate v., fig. 4.

1867. Palceozamia {Podozamitcs) viorrisii Tate, Quart. Journ.
Geol. Soc, vol. xxiii., p. 145, pi. v., fig. 4.

Frond pinnate ; rachis broad and woody, bearing two rows of
pinnae. The pinnye are attached either at right angles or at a very
wide angle to the rachis ; in shape like the head of a spear, rounded
at the base and tapering to an acuminate apex. Veins numerous
and parallel.

A portion of Tate's type-specimen from Geelhoutboom (Museum
of Geological Societv No. 11,108) of this species is figured on pi. v.,
fig- 4-

This solitary specimen may, as Tate believed, represent a well-
marked type, distinct from Zaiuites recta, but it is by no means
impossible that it may be a fragment from the basal portion of a
frond of that species. It may be compared with Otozamites
acuminatus (L. and H.) " from the Yorkshire Inferior Oolite, with
which it agrees in the position and form of the segments, but the
resemblance is hardly close enough to be regarded as a mark of
identity. Some of the specimens from the Potomac formation of
America, figm^ed by Fontaine as Nageiopsis ocata j are very similar
to Zamites morrisii. In the true Narjeiopsis the leaves are arranged
spirally as in the recent conifers Agathis and Podocarpus, whereas in
Zamites the segments are given off from the rachis in two rows.

Plate v., fig. 4.

A portion of the type-specimen is shown in the figure. The stout
woody axis bears two lateral rows of short acuminate pinnae, which
in venation and apparently also in their manner of attachment to the
rachis agree with Zamites recta.

Zamites africaxa (Tate).

[Possibly identical with Zamites ruhidgci (Tate)] .

Plate v., fig. 5.

1867. PcdcBOzamia (vel Pterophnllum) ufricana, Tate, Quart. Journ.
Geol. Soc, vol. xxiii. p. 145, pi. v., fig. 5.

The type-specimen of Tate (No. 11,110, Geological Society's

* Seward (00), pi. ii., fig. 1, pi. vi., tig. 1. f Fontaine (SO), pi. IxxviL, Ixxx.

26 Annals of the Soiitli African Museum.

Museum), is represented in pi. v., fig. 5 ; the original figure shows
only a single pinna.

If the specimen had not been named by Tate, I should have
probably not ventured to record it under a specific name. The
imperfect Hnear pinnae shown in fig. 5 are, no doubt, segments of a
pinnate cycadean frond similar to Zamites hiicliianus (Ett.) ; it is by
no means unlikely that the specimen named by Tate Palaozamia
rubidgei, may represent the apical portion of a frond of which the
fragments referred by him to Pakeozamia africaua, are the longer
and more spreading segments of the lower or median region.

Plate v., fig. 5.

(Museum of the London Geological Society, No. 11,110.)

An imperfect specimen from the Sunday Kiver, Geelhoutboom,
showing portions of five parallel pinnae, the longest of which reaches
a length of 13-5 cm., with a breadth at the lower end of 1 cm.,
tapering gradually towards a narrow acuminate tip. The veins are
represented as prominent ribs. The dotted outline in the figure
indicates that a portion of the rock is not represented.

Zamites rubidgei (Tate).
[Possibly identical with Zamites africaua (Tate).]

Plate v., fig. 3.

1867. Palceozamia rubidgei, Tate, Quart. Journ. Geol. Soc,
vol. xxiii., p. 145, pi. v., figs. 3a, 3b.

Frond pinnate, bearing two lateral rows of linear pinnte. The
pinnae are confluent and decurrent on the rachis ; veins numerous
and parallel.

The type-specimen of Tate (Museum of the Geological Society,
No. 11,109) is refigured on pi. v., fig. 3. I have not met with any
examples of this type in the collection recently received from Cape
Colony. It is possible that the discovery of more perfect material
may render it advisable to substitute for Zamites some other generic
name, but for the present I prefer to make use of that designation.
The fragment may be compared with such a fossil found as Zamites
buchianus (Ett.) of Wealden age,''' as described from England,
Germany, Japan, and elsewhere. The veins are parallel and, so far

' Seward (!to), p. 7'.), pi. iii., iv., viii. A. Nathorst (DO), pi. ii., iii., v. .

Fossil Floras of Cape Colony. 27

as one can see, not connected by transverse or oblique anastomoses,
but in other respects Tate's species agrees closely with Ctenis falcata
L. and H. A very similar form of leaf is represented also by a
specimen in the York Museum from the Inferior Oolite beds of East
Yorkshire ; but until we obtain more material Tate's species cannot
be more fully defined or compared with other types. The specimen
probably represents a portion from the apical region of a frond of
similar form to the Wealden species Zamitcs huchianns (Ett.).

Genus NILSSONIA.

NiLSSONIA TATEI Sp. nOV.

Plate IV., figs. 1, 2.

The specimens (291c, 293c), on which this species is founded, are
almost too fragmentary to name, but the portions of fronds repre-
sented in figs. 1 and 2, with an imperfectly preserved larger specimen
(293c), enable us to obtain some idea as to the general form of the
complete leaf.

Frond consisting of a prominent woody axis bearing a lamina
dissected into long band-like segments of unequal breadth, separated
from one another by fairly broad sinuses reaching to the rachis.
The lamina appears to have extended on to the upper face of the
rachis, and was probably continued over it as in the typical Nilssonia
fronds ; the segments are traversed by numerous parallel veins which
were single or forked close to their origin from the axis. The species
with which the Uitenhage fragments exhibit the closest similarity
have been described under the generic name PtcrophylliLm, but there
appears to be enough evidence in the present case to assign the
specimens to Nilssonia.

Among Indian forms, the following may be compared with
Nilssonia tatci : —

Pterophyllum livinceps 0. and M.,'''- P. rajmalialense Morr., t
P. carter ianum 0\di.;l P. morr isian urn Old., § P. falconerianum
Morr., II and P. medlicotianwm O. and M.*i Other similar species
are Pterophyllum kchncrscnianum Heer,"''' a Jurassic Siberian type,
Pterophyllum aq^tah Brongn., I f as figured by Nathorst from the

* Oldham and Morris (63), pi. x. t il>i<^; pl- xiii., fig. 4.

+ Ihid., pl. XV., fig. 4. § Ihid., tig. 1.

II Ihid., fig. 2. *\ Ihid., fig. 3.

** Heer (77), pl. xxv., fig. 5. tt Nathorst (7'J), pl. xv.

28 Annals of tlic Soutli African Museum.

Ehffitic of Scania, another European Eluctic species, Nilssonia
acuminata Goepp.,* and a Queensland fossil described by Shirley as
Pterophyllum quadriflorum.l

Plate IV., figs. 1, 2 (291c).

These fragments illustrate the characters referred to above. The
preservation is far from perfect, and it is possible that the reference
to Nilssonia is incorrect. In the lower right-hand segment of fig. 1
the lamina extends up to the middle of the rachis, and in this respect
agrees with species of Nilsso)iia.

Other Specimens : 293c, part of a frond 8 cm. long. The lamina
is much torn and broken, but there appear to be entire segments
broader than those represented in pi. iv., figs. 1, 2. The veins
are clearly preserved ; in the low^er part of the specimen they spring
from the rachis approximately at right angles, in the apical region
being given off at a much smaller angle ; they are approximately
1 mm. apart, rather less numerous than in the specimens repre-
sented in figs. 1, 2, pi. iv. The lamina, at least in the apical
region, appears to be practically continuous over the axis of
the leaf.

Also specimens of Araucarites scales.

Genus CYCADOLEPIS Saporta.

We owe this generic designation, to Saporta, :[ who made use of it
for fossils which he regarded as bud-scales of cycadean fronds.
In Part II. of the British Museum Catalogue of the Wealden
Flora § I extended this term to include " Scale-like leaf structures
of cycadean plants, varying considerably in form and including
detached petiolar bases, bud-scales, &c., also isolated carpellary
or antheriferous scales which exhibit no trace of ovules or pollen-
sacs."

We might add to this designation, scale-like structures which may
have served as a protective covering to young and unexpanded
vegetative leaves or fertile shoots.

* Schenk (HT), pi. xxxiii. f Shirley (1)8), p. Ki, pi. xix. and xxiv.

I Saporta (7-")), p. 200. 5^ Seward (Uo), p. <m;.

Fossil Floras of Cape Colony.

29

Cycadolepis jenkinsiaxa (Tate).

Plate IV., figs. 3-6.

Text-figure 2.

1867. Cycloptcris jcnkinsiana, Tate, Quart. Journ. Geol. Soc,
vol. xxiii., pi. vi., fig. 4, p. 146.

Scale-leaves more or less orI)icular in shape, reaching a length
of rather more than 12 cm. and a breadth of almost equal
dimensions ; attached to the supporting organ by a broad

^

y^jf:,iiS/4;i':/- :

i' ■'!

mm

Text-figure 2. (Nat. size.)
Cijcadolejns jenkinsiuna (Tate).

base. The lamina is traversed by a well-defined midrib consist-
ing of several parallel veins (fig. 5), but most of the veins spread
from the base towards the sides and distal edge, forking repeatedly
as in the Cyclopteris type of venation ; between and above the veins
occur numerous fine anastomosing and irregularly disposed Hnes
(figs. 4, 5) which present the appearance of the remains of a felt
of hairs which originally covered the whole lamina. The scale-leaf

30 Annals of the South African Museum.

often occurs folded on itself along the median line as if functioning
as a protective or stipular organ.

The specimen figured by Tate as Cyclopteris jenkinsiana is
undoubtedly identical with that represented in pi. iv., figs. 3-6.
An examination of his type-specimen in the Museum of the
Geological Society reveals certain imperfections in the published
drawing, although on the whole the figure is by no means inac-
curate ; the outline of the fossil is more ill-defined and irregular
than the drawing suggests, and under a low magnifying power the
surface resolves itself into an irregular network of fine lines, as
shown in fig. 5, pi. iv. (specimen 319c). Veins, like those seen
in fig. 6, may be detected here and there in Tate's specimen, but
for the most part they are too obscure to recognise ; the texture
of the leaf and the irregular fine lines over the surface present an
appearance as of some woolly covering over the whole lamina,
as shown also in fig. 4, pi. iv.

The specimen shown in text-figure 2 presents the appearance of
a flattened stem, but the surface features, as seen on the side
figured, are clearly those of Gi/cadolejns ; along one edge there
are distinct traces of veins following a vertical course, and from
them numerous veins spread in an obliquely ascending direction
towards the right-hand upper edge of the cast. A short distance
below the upper end of the specimen a transversely elongated
scar, s, marks the position of an appendage, which was possibly
another Cycadolcpis sheathing scale. Part of the surface is covered
by a thin mineral layer, h, which exhibits no definite texture. The
whole specimen is 8 cm. long by 4-5 cm. broad, and has a com-
pressed elliptical form in section : it represents, I believe, the cast
of a branch bearing several sheathing scale-leaves, and may well
belong to Bcnstcdtia. While admitting the absence of decisive proof,
I am inclined to regard Gycadolciyis as a bract- like organ borne on
the branches of Benstedtia and functioning as a protection to the
young and unexpanded leaves.

The specimens represented in pi. iv., figs. 3-6, present a fairly
close resemblance to some Wealden fossils described in 1895 as
examples of Saporta's genus Cycadoleins, and included in a sub-
section Eury-Cycadoleins, which was defined as follows : " Broadly
oval or orbicular scales, with the broadest portion frequently nearer
the distal than the proximal end; thick and fibrous structures." ■''■

Two examples of these scales are shown in text-figures 3 and 4.

* Seward (95), p. 'JG.

Fossil Floras of Cape Colony.

31

I regarded the English Wealden specimens as stout and leathery
scales which were originally attached to a cycadean stem, com-
paring them with the large persistent petiole bases of the recent
Macrozamia douglasi and other cycads. This comparison still
seems to me justified by the close resemblance between the
fossils and the scales of recent cycadean stems as well as with
such fossil scale-leaves or leaf-bases as those in Carruthers' genus
Fittonia.-' The English specimens usually show no trace of vena-
tion, but in the small example shown in fig. 3 there are several
clearly marked veins springing from the basal end. In the larger
specimen (fig. 4) the surface shows in places numerous irregular
wrinklings similar to those in Cycadoleins jenkinsiana. The curved
ridge stretching across the scale close to the base of the specimen

Text-figure 3. (Xat. size.) (British Museum Collection, Xo. V. 2131a.)
Cycadolepis from the Wealden of England.

drawn in fig. 4 shows the area of attachment and agrees closely
with a similar feature in the Uitenhage examples. In the African
specimens the lamina is often folded along the median line, a
character not noticed in the English forms, and the latter differ
in the apparently greater robustness and less leaf-like nature. On
the whole I am disposed to consider the Uitenhage fossils as
detached leaf-like bodies borne, possibly, on a cycadean stem
and serving as protective structures to leaves during the earlier
stages of growth. They may be compared with the stipules of
the Marattiacea^ among ferns, and with the much smaller scale-
leaves of modern cycads. It is true that existing cycadean plants
do not possess structui'es that function as sheathing bracts and bear
a close resemblance to Ci/cadolejns jenkinsiana ; a nearer analogy

* Carruthers (70), p. G90.

32

Annals of the South African Museum.

with the fossil scale-leaves is presented by the sheathing stipules
of such recent ferns as Marattia, An(iiopteris, Todea ; among
Palgeozoic ferns we have the so-called Aphlchia, which, in some
cases at least, acted as protective scales to unexpanded fronds. It
may be that the fossils referred to Cycadolepis should rather be
described as fern leaves, but I incline to the view that they are
cycadean for the following reasons : They correspond fairly closely

Text-figure 4. (J nat. size.) (British Museum Collection, No. V. 2799.)
Cycadolepis from tlie Wealden of England.

with the English specimens which I have little or no hesitation
in referring to cycads ; they occur in almost constant association
with the fronds of Zamites recta in the Uitenhage beds. "We know
that Mesozoic cycads possessed more fern-like characters than are
found among their modern representatives, and it would not be sur-
prising to find that some types produced scale-leaves comparable
with the AiMohia of extinct species of ferns and the stipular
appendages of certain existing genera.

Fossil Floras of Cape Colony. 33

Plate IV., tig. 3 (315c, 316c).

This specimen has a length of 7'0 cm.; it is not quite complete
.at the distal margin (lower end of drawing) ; its greatest breadth is
7 cm., but the edges are slightly torn and imperfect. At the upper
end, as seen in the figure (the base of the specimen), there is a
straight edge to the lamina, which no doubt marks the original place
of attachment. The lamina is traversed by numerous spreading
veins, as in a true Cijcloptcris, but the median portion is occupied
by more numerous parallel veins which form a fairly distinct and
broad midrib. Over part of the surface the veins are obscured
by the presence of fiexuous and irregularly anastomosing fine
lines. The resemblance to Tate's type-specimen is very close, as
regards shape, the presence of a midrib, and the surface texture-

Fragments of Zamites fronds and rachis with Araucarian scales
and seeds occur on the back of this slab.

Plate IV., fig. 4 (280c).

A scale similar to that shown in fig. 3, but rather less and with
a more complete base (upper end of drawing). The well-marked
slightly curved line at the termination of the midrib may be
regarded as the broad attachment-area by which the scale was
articulated to its supporting organ. The outline is torn and very
indistinctly defined ; the surface features agree exactly with those
in Tate's type-specimen.

Plate IV., fig. 6 (311c).

A scale 12 cm. long, torn across close to the base. The lower
margin is clearly defined, but the distal end and upper edge are
torn and incomplete ; the whole is traversed by veins, which are in
places hidden by the presence of the fine hair-like markings shown
between the ribs in fig. 5 ; the veins run parallel to the upper edge,
which is approximately in the position of the midrib, on which the
whole lamina was evidently folded as if clasping and protecting
some immature organ. The drawing, therefore, probably represents
a large scale-leaf folded along its axis.

Plate v., fig. 5 (319c).

A portion of a scale-leaf enlarged three times to show the veins
as prominent ribs and the delicate carbonaceous lines forming an
irregular network in the intercostal areas.

3

34 Annals of the South African Museum.

Other Sjjeciiuens : A, a specimen 7 cm. long by 6-5 broad,
convex in section, and evidently a portion of a broad scale-leaf
folded on itself ; veins clearly defined. 265c, a good example of
the basal portion of Cijcadolepis {Zamitcs on the reverse side).
305c, 313c, 314c, a specimen about 10 cm. long showing the
/eins over part of the sm-face. 313c, with Zamitcs pinnae.
320c, a good impression of part of a Cycadoleins showing the
Hne of attachment at the base and the spi-eading veins. The
broad base of the midrib is clearly preserved. A fragment of
Onychiopsis on the reverse side. 321c, with Zamites and ? Ben-
stcdtia. 324c, 326c {cf. 305c), 335c, and Zamitcs, 336c, 371c.

Genus BENSTEDTIA.

In 1862 Mr. Mackie " described and figured a fossil stem from
Lower Greensand rocks in Kent, which Konig named Draccena hcn-
stedtii. This name was adopted by Morris i and by Mantell. In
1868 Carruthers \ expressed the opinion that the fossils discovered
by Mr. Bensted presented a closer resemblance to Pandanus than to
Draccena, but Gardner § afterwards alluded to the specimens as
possibly cycadean. An examination of Mackie's original specimen
in the British Museum, and of other examples from Lower Green-
sand and Wealden rocks, led me to adopt the view that Konig's
DraccBua is in all probability the stem of a cycad. In most recent
cycads the surface of the stem is encased in an armour of persistent
leaf-bases, but in Zamia {e.g., Z. skinneri Warsz., Z. loddigcsii
Miq., &c.) and in Cijcas siamensis Miq. the leaf-bases are replaced
by a corky investment bearing numerous oval protuberances and
irregular transverse wrinklings. This close resemblance between
the fossil and recent stems caused me to institute a new generic
name, Benstedtia,\\ to be applied to such fossil stems as agree in
surface features with cycadean trunks of the Zamia type. Ben-
stedtia is defined as follows : —

" Stems having the surface marked by irregular and interrupted
grooves and broader ridges running transversely, with occasional
small elliptical protuberances irregularly disposed on the surface of
the stem. No distinct leaf-scars ; branch -scars may be present, and
in addition to smaller lateral branches, a bifurcation of the stem

* Mackie (02), pi. xxii. f Morris (54), p. 8.

+ Carruthers ((JH), p. 154 (footnote).

§ Gardner (8(j), p. 201. || Seward (96).

Fossil Floras of Cape Colony. 35

may be indicated by the converging upwards of the transverse Hnes
on the surface of the stem."

Since the publication of my paper in which the genus Bcnstedtia
was suggested, Professor Fliche, of Nancy, has described some casts
of stems from the Lower Cretaceous of France under the name
Conifcrocaidon.''- The fossils are, I believe, identical with those
referred to Bcnstedtia. Professor Fliche draws attention to the
resemblance as regards surface features to stems of Araucaria of the
section Colyinhcea, but does not suggest a comparison with Zamia.
Without anatomical evidence it is impossible to say whether the
fossil stems are cycadean or coniferous, but the resemblance to
Zamia is at least as striking as their similarity to Araucaria, and
in case of doubt a non-committed term such as Bcnstedtia is
preferable.

Bexstedtia sp.

(Cf. Coniferocaulon cohjmhea'forme Fliche.)

Plate v., fig. 2.

Text-figure 5.

The Uitenhage specimens referred to this genus are very similar
to those previously figured from Maidstone t and from the Sussex
coast. I They include partially flattened cylindrical casts, with
occasional branch-scars, and portions of much broader stems like
that represented in fig. 2, pi. v.

The collection of Uitenhage plants in the Museum of the
Geological Society contains a few imperfect specimens like that
shown in fig. 2, pi. v., which Tate identified as probably cycadean.^
I agree with his determination, and, while admitting the absence of
proof, am disposed to consider the specimens as casts of some
cycadean plant. The frequent association of these stems with the
fronds of Zamitcs recta is a point in favour of the supposition that
they represent portions of the same plant ; possibly the large Cyca-
dolepis scale-leaves may also have been borne on Bcnstedtia stems.
This is, however, merely a suggestion, in support of which I am
unable to furnish anything of the nature of proof.

The stem casts figured by Fliche || from Lower Cretaceous rocks
of France (La Haute-Marne) as Conifcroccudon colymheaforme

* Fliche (00). f Seward (96), pi. xiv. \ Seward (95), pi. xii.

§ Tate (67), p. 147. II FHche (GO), fig. 1-3.

36

Annals of the South African Museum.

appear to be identical with those from South Africa ; the surface
features are the same, and in both cases branch-scars occur. It is
possible, as Fliche suggests, that the stems are those of an
Araucarian plant, but whether their determination as coniferous or
cycadean is more likely to be correct must be left undecided.

Plate v., fig. 2 (292c).

A portion of a large piece of bark, 10 cm. by 9 cm. The axis
traversing the surface obliquely — as shown in the drawing — is prob-

Text-figdre 5.
Benstedtia sp. 387c. (Nat. size.)

ably a foreign body, and may be the rachis of Zamites recta. The
surface characters as shown in fig. 2 appear to be exactly of the type
described in the Enghsh specimens of Benstedtia.

Text-fig. 5. A flattened cast, 8-5 cm. long and 27 cm. broad,
showing one branch-scar, s. The cast is for the most part covered
with a layer of bark exhibiting the characteristic tuberculated and

Fossil Floras of Cape Colony. 37

wrinkled surface of Benstedtia. Very similar to Fliche's specimen
represented in his fig. 2. On the face of the cast shown in the
drawing the bark is repi'esented by a very small piece — b, exhibiting
the Benstedtia texture.

Other Specimens : 297c, an almost flat impression, 12 cm. long,
with some carbonaceous matter on the surface, bearing numerous
tubercles, reminding one of the transversely elongated lenticels on a
piece of birch bark. 338c, with fragments of Cladophlehis broioniana
and a coniferous twig. 364c, with Zauiitcs, 379c the cast of a stem
32 cm, long, showing a single branch-scar ; one side is covered with
a layer of bark similar to that shown in pi. v., fig. 2, resting on the
woody substance of the stem. 380c, a narrow branching cast,
probably referable to Benstedtia ; 381c, 382c, 383c, 384c, 386c.

CYC ADOPHYTA ?

Genus CAEPOLITHES Sternberg.
Carpolithes sp.

The collection includes a single seed (388c) of oval nut-like form,
1-3 cm. long by 9 mm. in breadth. In shape it represents that of a
recent cycad rather than the seed of a conifer, and may be included
under Gymnospermae as a doubtful representative of the Cycadophyta.

Gkoup coniferales.

Family AEAUCAEIE^.

Genus ARAUCAEITES Presl.
Aeaucarites kogebsi * sp. nov.

Plate VI., figs. 4-7.

Cone-scales broad, reaching a breadth of 3 cm. ; distal margin
straight, prolonged into a short median appendage (fig. 4), and

* After my friend, Mr. A. W. Rogers, of the Geological Commission of Cape
Colony.

38 Annals of the South African Museum.

bending down into a thin lateral wing, as in the scales of Araucaria
cookii and other recent members of the Eutacta division of
Araucaria ; surface of the scales characterised by the occmTence of
forked ribs (vascular bundles), which follow a more or less vertical
course ; each scale bears a single obovate seed, terminating distally
in a bluntly rounded base. No trace of a "ligule."

The collection includes several examples of cone-scales like those
represented on pi. vi., figs. 4-7, which agree so closely with the
cone-scales of Araucaria cookii and other recent species of Araucaria
that I have no hesitation in referring them to the genus Araucarites.
They are practically identical wath the recent type of ovuliferous
seeds characteristic of the Eutacta section of Araucaria. There is
one specimen in the Museum of the Geological Society (11,117)
collected by Atherstone in 1858, which is no doubt the example
referred to by Tate as " the under surface of the base of a cone," '''
and doubtfully determined as cycadean. This specimen represents
several broad scales adhering together in their original position.
Among fossil forms there are several Mesozoic species which pre-
sent a close resemblance to Araucarites rogersi, but the agreement
is hardly such as to justify the inclusion of the African cone-scales
in any previously recorded species. Araucarites macropteris f and
A. cutcliensis,l described by Feistmantel from the Eajmahal and
Jabalpur series of India, represent closely allied types. Araucarites
loyomingcnsis Ward,§ from the Neocomian of Dakota, affords
evidence of the occurrence of a similar species of the genus in
North America. Among English species Araucarites hrodiei,\\
A. pMllipsiiyM A. splicerocarpus ;•'■'' described by Carruthers from
Jurassic strata, afford other examples of broad Araucarian cone-
scales ; and a specimen figured by Fric and Bayer from Lower
Cretaceous rocks of Bohemia as " cf. Araucaria bohcmica Vel.," f I
may also be mentioned as evidence of the existence of this type of
AraiLcaritcs in Europe.

In the absence of more perfect material it is probably better to
include the smaller scale represented in fig. 5, pi. vi., with the larger
form under one specific name.

* Tate (07), p. 147.

t Feistmantel (77"), pi. viii., figs. 9-12 ; (79), pi. xiv., xvi.

I Feistmantel (79), pi. xiv., xvi.

§ Ward (99), pi. clxiii., figs. 1-6, 8, 9.

II Carruthers (G9-), pi. ii., figs. 1-6.

•f IhiiL, pi. ii., fig. 7-9 ; Seward (00), pi. x., fig. 4.
»* Carruthers (OC.) (71). If Fric and Bayer (01), fig. 50.

Fossil Floras of Cape Colony. 39

Plate VI., fig. 4 (2980).

Tliis scale — 3 cm. long — shows clearly the impressions of the
vascular bundles over the area originally occupied by the seed. The
thin lateral margins are partially preserved, but in the perfect scale
the breadth was no doubt sliglitly greater. In the middle of the
distal edge the scale shows part of a projection similar to the longer
outgrowth seen in Araucaria coohii, the scales and seeds of which
agree closely with those of Araucaritcs rogersi.

Plate VI., fig. 5 (298c).

A much smaller and less perfect scale, similar to Araucarites
cutcliensis, as figured by Feistmantel from the Upper Gondwanas of
the Madras Coast,* and to some scales of Araucaritcs pliiUipsi Carr.f

Plate VI., fig. 6 (295c).

Two imperfectly preserved overlapping scales, each showing the
outline of a seed. The two scales are so closely pressed together
that it is almost impossible to trace the boundary between them,

Plate VI., fig. 7 (300c).

This specimen show^s more clearly the broad membranous wing.
Part of the projection from the distal margin is preserved, a depres-
sion in the middle line of the scale indicating the position of the
seed.

Other Specimens : 301c, 304c (with fragments of Zaniitcs), 306c,
307c, 308c (with Zaniites, &c.), 319c (with good impression, on the
reverse side, of Ci/cadolepis), 322c, 330c, 331c, 337c.

CONIFEEALES OP DOUBTFUL POSITION.

Genus TAXITES Brongniart.

This artificial genus is convenient for the inclusion of fragments
of vegetative twigs, presumably of conifers, resembling those of
Taxus, Sequoia sempcrvircns, Torreija, and some other types of
recent Coniferae, in having narrow linear leaves traversed by a single
midrib and disposed in a two-ranked arrangement. Many of the

* Feistmantel (79), pi. xiv. f Cf. Seward (CO), pi. x., tig. 4.

40 Annals of the South African Museum.

fossil fi'cagments referred by authors to the genus Palissya and to-
Sequoia have the vegetative characters of tlie comprehensive genus
Taxitcs.

Taxites sp.

Plate VI., fig. 15.

The fragment represented in fig. 15 is too small and incomplete
to refer to a distinct species ; twigs of this type are exceedingly
common in Mesozoic Floras, and, in the absence of reproductive
organs, cannot be placed with any degree of certainty in a particular
family. Similar coniferous shoots are recorded by Feistmantel from
India as Taxites ^;/a727«s ''' and Palissya (?) jahalj^Jirensis ; i the
British type, Taxites zamioidcs,\ of Inferior Oolite age, is another of
the numerous examples that may be compared with the Uitenhage
specimen.

Plate VI., fig. 15 (Y).

A fragment showing numerous linear leaves attached by a base
which is decurrent on the axis of the shoot. The leaves appear to
have a single median vein.

Associated with fragments of Cladoplilebis denticulata forma
atherstonei.

The only other specimen in the collection is 176d.

Genus BEACHYPHYLLUM Brongniart.
Brachyphyllum sp.

Plate VI., figs. 13, 18.

The scraps of coniferous twigs preserved in the Uitenhage plant-
beds are too fragmentary and imperfect to refer with confidence to
any species. § Tate mentions the occurrence of portions of a
coniferous stem, which are no doubt identical with that represented
in fig. 13, " closely allied to Athrotaxitcs indicus Old." || This type
of stem is exceedingly abundant in Wealden and Jurassic rocks ; it
is recorded by many authors under the genus Brachyphylluvi, while
others adopt the names Echinostrobus or AtJirotaxitcs. Feistmantel's

* Feistmantel (79), pi. xiii.-xv. f Feistmantel (773), pi. ix., x-

+ Seward (00), pi. x., fig. 5. § Tate ((17), p. 147.

11 Oldham and Monis ((J.'J), pi. xxxii., fig. 8.

Fossil Floras of Cape Colony. 41

Echinostrobus rajmahalensis may bo veferrecl to as one of the many
similar types of Coniferous shoots, also the widely spread species
Brachyphyllum mainillare Brongn., but without more material
specific identification is hopeless.

Plate VI., tig. 13 (299c).

The figure represents a portion of the specimen ; the branch was
clothed with crowded and spirally disposed leaves, which were
probably of the thick and fleshy Brachypliylliim type.

Associated with Araucarites scales and with Zaviites recta.

Plate VI., fig. 18 (348c).

A small piece of a branched twig showing a well-defined reticulum
marking the spaces between the crowded leaves or leaf-cushions.
Each mesh of the reticulum is occupied by a flat area divided intO'
an upper and a lower region, and showing in the upper third of the
median line a patch of carbonaceous material. The whole (fig. 18)
presents a superficial resemblance to the leaf-cushion and leaf-scar
of a Leiyidondendron.

The portion figured has a length of 3*5 mm.

In addition to fragments like that represented in fig. 13, pi. vi., a
few smaller examples of coniferous twigs occur (354c, 355c) of the
type shown in fig. 2. These smaller pieces show the apices of the
appressed and thick leaves projecting slightly from the axis of
the branch. A comparison may be made between these more
slender shoots and the European Wealden species Sphemlepidmm
kurrianum (Schenk), but it would be rash to suggest specific
identity.

Genus CONITES Sternberg.

CONITES Sp. a.

Plate VL, figs. 2, 2a.

The two small and somewhat obscure cones shown in fig. 2 (310c)
appear to be attached to a slender branched axis, but they are not
in organic connection with the adjacent twig. It is impossible to
determine the nature of these cones, whether male or immature
female ; the small dimensions suggest a comparison with the male
flowers of recent Conifera3, but, on the other hand, the comparatively

42 Annals of tJic South African Miiscmn.

broad character of the cone is a point in favour of a comparison with
such small, presumably female, cones as those of the Wealden
Splienolcpidium sternhcrgianuvi (Schenk),''' Ceratostrobus echinatus
Vel.,f of Lower Cretaceous age, and some small cones described by
Fontaine from the Potomac formation:!: as " aments of conifers."
The cone-scales as shown in fig. 2a are imperfectly preserved, but
there aref indications of a central keel or projection on their
■expanded distal ends.

CONITES Sp. /3.

Plate VI., figs. 1, la, 3.

These cones may be specifically identical with the slightly broader
form represented in fig. 2, but their relatively longer and narrow
form suggests that they are male flowers.

Plate VI., figs. 1, la (253c).

The cone represented natural size in fig. 1 shows numerous
spirally disposed, distally expanded scales with traces of a small
central umbo (fig. la), which may well be sporophylls bearing
pollen-sacs.

Plate VI., fig. 3 (298c).

A cone seen in longitudinal section, showing the sporophylls
attached at right angles to the axis of the flower.

Other Speciniois : 250c, 251c, 252c.

CONIFEEOUS WOOD.

In addition to the stem casts referred to the genus Benstcdtia, the
collection includes a few specimens of what is probably true coni-
ferous wood ; but without petrified examples it is impossible to
attempt an accurate determination.

Specimen x. Tliis specimen represents a sandstone cast, 19 cm.
long, and at least 8 cm. in diameter. The contour of the surface is
the chief point of interest ; the wood of the stem or root was con-

* Schenk (71), pi. xxxviii., fig. 13. f Velenovsky (85), pi. xi., figs. 13, 15.

\ Fontiiino (si)), pi. cxxxvi., fig. 2.

Fossil Floras of Cape Colony. 43

siderably swollen at intex'vals ; the cast presents the appearance of
a transversely corrugated axis with l)road, transverse swellings and
intervening grooves. There are two branch-scars. The irregularity
in the diameter suggests a comparison with the root wood of recent
conifers. Somewhat similar casts are recorded by Goeppert •■' from
the Quadersandstein of Silesia as Ci/lindrites spongioiclcs.

Specimens 32c, 33c, 34c, 36c, probably represent portions of coni-
ferous wood, but in the absence of petrified tissues an accurate
■determination is out of the question.

PLANT OF DOUBTFUL POSITION.

Plate VI., fig. 14.

The specimen represented in fig. 14 (332c) consists of two thick
bodies, consisting of a short stalk terminating expanding into a
semicircular head. In the lower part of the expanded head there
is a well-defined depressed area, bounded below by a straight line
and by a curved line above, which may be the scar of some
appendage. The surface is irregularly pitted.

The collection includes a few other isolated bodies like the two
shown in the drawing (f.{/., 326c).

Some detached scales figured by Fric and Bayer f from Lower
Cretaceous rocks of Bohemia as Dammara horcalis Heer bear a
resemblance to fig. 13, but I must leave them as portions of a plant
of uncertain position.

III.— CONCLUSION.

The following list includes all the species and genera recognised
in the Uitenhage Flora ; the localities from which the specimens
were obtained are added after each name : —

FiLICALES.

Onychlopsis mantelU (Brongn.) Bezuidenhout River, Witte River. (Both

localities on the farm Geelhoutboom,
now the Roman Catholic Mission Sta-
tion, Dunbrodie.)

* Goeppert (41), pi. xlvi., xlviii., p. 11.').
t Fric and Bayer (01), p. it5, tig. 47.

44 A7inals of the South African Museum.

Cladoj^hlehis hroiniknia (Dunk.) Bezuidenhout Kiver.

Cladoplilehis dcuticulata (Brongn.) forma

atherstonei Herbertsclale, Bezuidenhout River.

Splienoiiteris fittoiil Sew Bezuidenhout River.

Sphcnopteris sp Sunday River, Dunbrodie.

Tccniopteris sp Herbertsdale, Bezuidenhout River.

CyCADOPHYTA.

Zamites recta [T&ie) Bezuidenhout River.

Zamites morrisii (Tate).
Zamites africana (Tate).
Zamites ruhidgei (Tate).

Nilssonia tatei sp. nov Bezuidenhout River.

Cycadolejns jenkinsiaiia (Tate) Bezuidenhout River.

Benstedtia sp. (cf. Coiiiferocaiiloii coliim-

heceforme Fliche) Bezuidenhout River.

Carpolithes sp Bezuidenhout River.

CONIFERALES.

Araucarites rogersi sp. nov Bezuidenhout River.

Taxites sp Herbertsdale.

Brafhyphijllum sp Bezuidenhout River.

Conites sxi. a.\ „ ■, , , -n-

,-^ . '^ ^ L Bezuidenhout River.

Conites sp. /5. i

Coniferous Wood Bezuidenhout River.

Planta incertcc sedU Bezuidenhout River.

Onychioi)sis mantclll (Brongn.). This species, although probably
not absolutely confined to rocks of Wealden age, is essentially a
characteristic member of floras belonging to that epoch. It has a
wide geographical distribution in Europe, and occurs in North
America and Japan. Onychioims mantclli must have been one of
the most abundant plants in the vegetation of which fragments are
preserved in the Uitenhage beds.

Cladophlchis hroioniana (Dunk.). As already pointed out, it is
almost impossible to determine fronds of this type with accuracy,
])ut the specimens with which the South African specimens appear
to be identical is a Wealden fern of common occurrence in Europe
and in the Far East. The distinction between Wealden and Inferior
Oolite species is in many cases by no means well marked, and this
form of frond, taken by itself, cannot be regarded as a trustworthy
index of the Wealden age of a flora.

Cladophlehis denticulata (Brongn.) forma atherstonei. This repre-

Fossil Floras of Cape Colony. 45

sents another very abundant and widely distributed type of frond
met with in rocks ranging in age from Ehyctic to supra- Wealden.
Taken by itself this almost ubiquitous form of fern does not atibrd
conclusive testimony of age, but it is chiefly characteristic of the
Upper rather than the Lower Jurassic series, and of Lower Cre-
taceous rocks.

Sphcnopteris fittoni Sew. A less widely spread type and
a safer criterion of age, indicating Wealden rather than a lower
horizon.

Sphcnopteris sp. Too fragmentary a specimen to serve as a safe
guide.

Tceniopteris sp., cf. T.arctica (Heer.). Another common type, but
on the whole nearer to a species of Lower Cretaceous age than to
Jurassic types of Tcenioptcris.

Zamites recta (Tate). The fronds of this species are very similar
to those of WiJliamsonia gigas (L. and H.), a characteristic Jurassic
plant. The separation of pinnate cycadean fronds of the form
represented by Zamites recta into well-defined species is hardly
possible; but whether Tate's species be distinct, or identical with a
European type, it undoubtedly agrees more closely with Jurassic
than with Wealden cycadean fronds.

Zamites morrisii (Tate). The type-specimen of Tate is the only
example so far obtained ; it may be a distinct species, but it is by no
means unlikely that it may prove to be identical with Z. recta. It
throws no further light on the question of age.

Zamites africana (Tate). Too fragmentary to be used as evidence
of age, but resembling a W^ealden species, Zamites huchianus (Ett.),
perhaps more closely than a Jurassic type.

Zamites ruhidgei (Tate). Fronds are known of both Jurassic and
Wealden age which bear a close resemblance to the single specimen
on which this species was founded by Tate.

Nilssonia tatei sp. nov. On the whole species of Jurassic age
from Indian and European localities exhibit the closest resemblance
to this form.

46 Annals of the South African Museum.

Ci/catlolepia joikin-'^iaiia (Tate), and The scale-leaves and stems from English
Bcnxtedtia sp. and Wealden beds referred to in the

account of these Uitenhage fossils show
a very close resemblance to the African
specimens.

Araucaritcs rocjcrsl sp. nov. The genus Araucaritcs was widely
distributed as a member of Jurassic floras, but its occurrence also in
Wealden beds renders it of uncertain value as an index of age.
Such forms of Araucarian cone-scales as resemble Araucaritcs
rogersi most closely are from Jurassic horizons.

It seems clear that the Uitenhage plants include types in part
cliaracteristic of Wealden and in part indicative of Jurassic floras.
On the whole there is a balance of evidence in favour of a Wealden
horizon. The comparative paucity of species and the fact that
several forms are represented by small fragments render conclusions
as to age somewhat difficult ; but I have little hesitation in stating
that the flora exhibits moi'e well-defined points of contact with
plants of Wealden age than with older floras. The impression,
given by the occurrence of Wealden species with others more nearly
allied to Jurassic types is that this flora from Cape Colony, repre-
sented by a small number of species, marks a phase of Mesozoic
vegetation on the boundary of the Jurassic and Wealden periods.
Between Wealden floras, especially such as are known from the
lower beds of the formation, and Oolitic floras, there exist but minor
differences. It was during the period immediately succeeding that
represented by the English Wealden strata that the plant-world
experienced a striking change. Before this change took place there
appears to have been a long period during which but little alteration
occurred in the composition of the Mesozoic vegetation.

Fossil Floras of Cape Colony.

47

ll.—STOBMBERG FLORA.

Plates VII.-IX., XI., figs. 2, 3.
Text-figures 6 and 7.

I.— INTEODUCTION.

The plants included in this section were obtained from the following
localities : Maclear and neighbourhood, Tina Eiver, Kenigha Eiver,
Matatiele, Edward's Hope, Cyphergat, Molteno, and Stormberg,
The most important account of the flora of this uppermost division
of the Karoo formation is that by Feistmantel," who refers to the
Stormberg beds as probably of Lias-Ehaetic age. The floras of other
regions — -Scania and Franconia in Europe, Australia, and the Argen-
tine Eepublic in the Southern hemisphere — with which the African
plants are most nearly allied are usually described as Ehastic in age,
and there is little doubt that the Stormberg flora should be regarded
as a member of the EliEetic vegetation. Feistmantel records the
following species from Indwe Eiver, Stormberg, Cyphergat, and
Molteno :—

Feistmantel (1889).
Equisetaceous stem fragment.

Sphenoi^terix elongate Carr.

Thinnfcldia odoiitopteroidcx (Morr.).)
Thinnfeldia trilohita (?) Johnst. |

Tteniopteris carriithersi T. -Woods. |
Tceniopteris daintreei McCoy. )

AntJiroplnjopds (?) sp.

Alethopteris sp. (cf. Aifpleniitm nehhense
Heer).

Podozamitc^ {Zeu(iop)lujllitcs) dongatuft \

(Morr.). I-

Podoznmites (Zeiigophijllites) sp. J

Sewakd (1903).
Equisetaceous stem fragment.

Stenopterix clougata (Carr.).
Thinnfeldia odontopteroides (Morr.

TaniopteriA carruthcrd T. -Woods.

Ghiroptens elongata (?) (Carr.).
Cladophlehls sp.

Phccnicopsis elongatd (Morr.).

Baiera schencki Feist.

I Baiera schencki Feist.
* Feistmantel (89).

48 A)i)ials of the South African Museum.

II.— DESCKIPTION OF SPECIMENS.

Geoup equisetales.

Genus SCHIZONEURA Schimper and Mougeot.

SCHIZONEURA KBASSERI Sp. nOV.

[Probably identical with Schizoneiira carrcrei Zeiller.]"
Plate IX., figs. 5, 6.

1900. ? Scliizonenra sp. Krasser, Denksch. K. Akad. Wiss. Wien,
Bd. LXX., p. 146, pi. iii., figs. l-3a.

Stems broad ; internodes long, reaching a length of over 20 cm.,
the internodal ridges crowded, pursuing either a straight course
across the node or occasionally alternating. Nodes shown as slight
depressions with indications of small leaf-trace protuberances.

Although it is impossible to give a satisfactory diagnosis of the
Stormberg Equisetaceous stems, I have ventured, for the sake of
convenience, to append a specific name, and in view of the probable
identity of the African plant with that described by my friend Dr.
Fridolin Krasser from China, I have designated the Stormberg type
Scliizoneura krasscri.

The fragments of broad Equisetaceous stems are too small and
imperfect to admit of a complete diagnosis : they are characterised
by the numerous internodal ridges which often pursue a straight
course across the nodal region and by their long internodes, which
reached a length of at least 7 cm. The small piece of ribbed stem
figured by Feistmantel f from the Stormberg beds shows broader
and more prominent ridges than the specimens represented in
figs. 5 and 6. The Scanian Ehaetic species Schizoneura hoerensis
His. I resembles the figured specimens in the narrow and numerous
ribs, and a piece of broad stem described by Zeiller § from Tonking
as Phyllothcca indica Bunb. is also very similar to the examples shown
in figs. 5 and 6.|| Specimens which may well be identical with the
South African species are figured by Krasser from China as Equi-
setaceous stems probably referable to Scliizone^ira,*\\ and by Schenk
.as Schizoneura sp. '■'•'' from Rhsetic rocks of Persia. Some of Feist-

* Zeiller (02=), pis. xxxvii. , xxxviii.
t Feistmantel (H9), pi. iii., fig. 9.

+ Nathorst (78), pi. x., figs. G-8. See also Nathorst (78), pi. i., figs. 1-4.
§ Zeiller (82), pi. x., fig. 1.
II See note. p. 75.
1i Krasser (00), pi. iii., figs. l-3rt.
*• Schenk (87), pi. viii., figs. 46, 48.

Fossil Floras of Cape Colony. 49-

mantel's figures ■■'■ of the stems of Schizoneura gonrhvanensis agree
very closely with figs. 5 and 6, pi. ix., and the fragmentary leaf-like
impressions with dark veins found in association with the African
stems are very similar to the Indian specimens of Schizoncura leaves.
The largest specimens of Schizoneura krasseri that I have seen
is one in the British Museum Collection (V. 2413) ; it consists of a
flattened stone-cast having a diameter of 10 cm. One node is
preserved, and an internode reaches a length of 20 cm.

Plate IX., fig. 5 (308d).
Several flattened impressions of broad stems ; the one figured has
a breadth of 6 cm., but is imperfect. The internode extends for a
length of 7 cm., but only one node is seen on the specimen. The
preservation of the nodal region is not sufficiently good to show the
structural details ; many of the ridges are seen to be continued in a
straight line across the node, but some show a distinct alternation.

Plate IX., fig. 6 (487d).

A portion of an imperfect broad stem showing a rather more
clearly defined nodal region with a few small and indistinct pro-
tuberances probably marking the position of leaf-traces. On the
reverse side of the specimen there are a few fragmeiits of what may
be leaves with broad parallel veins about 2 mm. apart ; possibly
these may be leaves of Schizoneura.

Other Specimens : B, good impression of stem showing a node and
indications of leaf-traces ; 104a, a piece of a narrower stem, 3 cm.
broad, showing a nodal construction; 481d, 484d (roots?); 485d, im-
pressions of stems, and roots? ; 488d (roots?) ; 489d, 494d, a well-
preserved specimen ; the surface of the ril)s shows fine longitudinal
striations, which no doubt indicate a woody structure ; 496d, ribs
rather broader.

Genus STEOBILITES Schimper and Mougeot.

Strobilites sp.

(Possibly the strobilus of ScJiizonenra Krasseri, the plant represented
in pi. ix., figs. 5, 6.)

The imperfectly preserved specimen shown in fig. 3, pi. ix. (486(1,
495d) is undoubtedly a strobilus bearing crowded appendages, pro-

* Feistmantel (80), pis. i., ii. ix., &c.
4

50 Annals of the South African Museum.

bably of the nature of distally expanded sporangiophores ; but the
preservation is unfortunately too imperfect to admit of any satis-
factory diagnosis. The cone itself is incomplete and does not show
the apical portion ; it is 6 cm. long and slightly more than 2 cm.
broad ; the surface is rough and uneven, but shows here and there
distinct indications of polygonal protuberances and a few small oval
bodies, which may be the terminations of sporangiophores bearing
■sporangia. The peduncle is represented by an axis 6 mm. broad
showing longitudinal striations on its surface. The form of the
€one, the appearance of the polygonal prominences and depres-
sions, and its association with stems like those shown in pi. ix.,
tigs. 5 and 6, lead me to regard the specimen as a strobilus of an
Equisetaceous plant, possibly Schizone^iva krasseri.

Group FILICALES.

FEENS OF UNCEETAIN SYSTEMATIC POSITION.

Genus THINNFELDIA Ettingshausen.

This generic name was instituted in 1852 for some specimens
from the Lias of Steierdorf and defined as follows : — *

" Eami teretes vel subalati. Folia disticha, alterna oppositave,
rhomboidea, ovalia vel lanceolata vel linearia, flabellatim vel
pinnatim venosa."

Ettingshausen's species Tliinnfcldia rhoviboidalis closely re-
sembles the South African forms, and represents a typical
example of the genus. The comparison made by the authors
of the genus with the New Zealand Conifer Pliijllocladus has
not been borne out by the investigation of more recently acquired
specimens.

In 1853 Andrae substituted Brongniart's genus Pachyptoris for
Thinnfeldia and spoke of T. rhoviboidalis as Pachyptcris thinnfeldia
And. f The genus Pachyptcris is defined by Brongniart as being
characterised by entire pinnules without veins or with a single vein,
and is referred by him to the ferns. Zigno's large and well-preserved
specimens from the Oolitic rocks of Italy placed in his genus
Dichopteris are no doubt generically identical with Pachyptcris,
but in the Italian examples I the pinnules are traversed by several

* Ettingshausen (52), p. 2. f Andiae (53), p. 43.

I Zigno (56), pis. xii., xiii., xv.

Fossil Floras of Cape Colony. 51

divergent veins. Species of Padiijptcris or Dlchoptcris in whicli
the ultimate segments possess spreading and forking veins bear a
marked resemblance to TJiinnfcldia ; it is probalilo that the two
genera are very closely allied.

Schimper, in the second volume of the " Trait6 de pal^ontologie
vegetale,"''' extends Ettingshausen's definition, and compares Thinn-
feldia with Pachypteris and Cycadopteris. In 1867 Schenk described
specimens of Thinnfeldia from the Khaetic beds of Franconia and
published figures of the epidermal cells and stomata ; he placed the
genus among the Cycads. f

Bornemann,:[: in 1856, published figures of the epidermis of
Scytophyllum hcrgeri Born., a plant bearing a close resemblance
to Thinnfeldia, which may be compared with the drawings repro-
duced in pi. ix., figs. 7 and 8. Solms-Laubach ;; notes the occurrence
of stomata on both sides of the lamina of Thinnfeldia as a character
which may possibly indicate that the genus may not be a fern ; in
other respects the fern-like nature appears to be clearly indicated.

Another genus — Ptilozamites of Nathorst — differs from Thinn-
feldia in the greater parallelism of the veins in the pinnules, but
in other respects there is a striking agreement between the two
genera, and it is doubtful if there are sufficient grounds for regarding
the slight difference in venation characters as more than a specific
distinction. A comparison of Nathorst's figures of species of Ptilo-
zamites from the Ehaetic of Bjuf :| and the Thinnfeldias of South
Africa represented in pis. vii. and viii. demonstrates the similarity
between the Northern and Southern types. Gutbier's Permian
species, Odontopteris rcichiana*' is another type wdiich exhibits
a very close resemblance to Tliinnfeldia. Another plant which I
venture to think should not be separated from Thinnfeldia has
been described and figured by Sellards from Permian rocks of
Kansas as Glenopteris simplex/'-''- In speaking of the characters
of this new genus Glenopteris the author notes its resemblance
to Loniatopteris and Cycadopteris, but does not refer to Thinnfeldia.
It is not clear that the institution of Glenopteris was a necessary
addition to palseobotanical nomenclature.

Similarly some of the Italian species of Dichoptcris of Zigno and
species referred by Saportaft to Loniatopteris and Cycadopteris may
well be generically identical with Ettingshausen's genus Thinnfeldia.

* Schimper (70), p. 494. t Schenk (67), p. lOo, pi. xxvii.

X Boraemann (o6), pi. vi., figs. 1-0. S Solms-Laubach (<J1), p. 140.

II Nathorst (78), pis. xii., &e. 'i Weiss (69), pi. i., figs. 3-9.

** Sellards (00), pi. xxxix. ft Saporta (73), p. 340.

52 Annals of the South African Museum.

Until we find well-preserved fertile specimens of Pachyperis, Ptilo-
zamites, Thinnfeldia, and other genera we cannot hope to define with
accuracy their systematic positions or relationships. One impression
derived from a comparison of Nathorst's Ptilozamites, of which there
is an exceedingly good series in the Stockholm Museum, with the
Southern Hemisphere Thinnfeldias is that the differences betw^een
the two genera have been exaggerated, and geographical separation
has been allowed too much weight in influencing the choice of distinct
generic designations. On the other hand, it is clear that certain
authors have imduly extended the application of Ettingshausen's
Tliinnfeldia .

The most satisfactory evidence so far published as to the systematic
position of Thinnfcldia has been furnished by Eaciborski in his de-
scription and illustration of a specimen of T. rhomboldalis from Jurassic
rocks of Poland : this author gives drawings of a leaflet bearing the
impression of what appears to be a circular sorus showing the
boundaries of individual sporangia, but the material is too imperfect
to afford any indication as to the nature of the sporangia.* In addition
to the evidence afforded by the Polish specimen I may mention a
specimen in the British Museum (V. 5950) in which the broadly
linear segments of a Thinnfeldia pinna show two rows of contiguous
polygonal or oval prominences, one on either side of the midrib,
which agree in size and shape with the sori of certain ferns ; but.
until we obtain sporangia we cannot speak with certainty as to the
precise nature of the fertile frond. My friend Mr. Arber, of Cam-
bridge, hopes to publish a more detailed account of the British
Museum specimen.

Thinnfeldia odontopteroides (Morris).

Plate VII., figs. 1, 7, 8, 8a ; Plate VIII., figs. 7, 8 ; Plate IX.,
figs. 7, 8; Plate XL, fig. 2.

Text-figure 6.

1845. Pecojyteris odontopteroides Morris, in Strzelecki's New South

Wales, p. 249, pi. vi., figs. 2-4.
1847. Crleichenites odontopteroides McCoy, Annals Mag. Nat. Hist.,.

vol. XX., p. 147.
1872. Pecopteris odontopteroides Carruthers, Quart. Journ. Geol. Soc,

vol. xxviii., p. 355, pi. xxvii., figs. 2, 3.

* llaciborski (94), pi. xx., figs. 1, 2, p. 20(5. See also Zeiller (00), p. 97.

Fossil Floras of Cape Colony. 53

1875. Pccopteris odontopUroidcs Crepin, Bull. Acad. E. Belgique,

vol. xxxix., p. 258 (plate).

1876. Thinnfcldia crassinervis Geinitz, PaUeont. Suppl., iii., p. 4.,

pi. i., figs. 10-16.
1883. Thinnfcldia odontoptcroides Tenison-Woods, Proc. Linn. Soc.
N. S. Wales, vol. viii., p. 103.
T. odontoptcroides xiiv. falcata, p. 107, pi. viii., fig. 1.

1888. Thinnfeldia odontoptcroides Szajnocka, Sitzh. K. Akad. Wiss.

Wien, vol. C. Abth. i., p. 10, pi. i., figs. 1-4.

1889. Thinnfeldia odontopteroides Feistmantel, Abh. K. I)ohm. Ges.

Wiss. Bd. III., Folg. vii., p. 62, pis. 1-3.

1890. Thinnfeldia odontoptcroides Feistmantel, Mem. Geol. Surv.

N. S. Wales, p. 101, pis. xxiii.-xxv., xxviii., xxix.
1892. TJiinnfeldia odontopteroides Jack and Etheridge, Geol. and

Palgeont. Queensland, pi. xvii., fig. 1.
1902. Thinnfeldia odontoptcroides Arber, Quart. Journ. Geol. Soc,

vol. Iviii., p. 2.

In addition to the references given in the synonomy, the following
may be added as types of Thinnfeldia which, may be either specifically
identical with TJiinnfeldia odontoptcroides or at least very closely
allied forms : —

Thinnfeldia rhomhoidalis Ettingshausen.'''

T. lancifolia as figured by Morris,! Szajnocha, | and Solms-
Laubach. §

T. tenuinervis Geinitz. ||

T. indica Feistmantel. 'i

Odontopteris viacrophylla Curran.**

Thinnfeldia falcata Tenison-Woods. I f

Glcichenia lineata Tenison-Woods. it

Frond bipinnate, rachis stout, occasionally bifurcating into two
linear pinnae inclined to one another at a small angle (pi. vii., fig. 1).
The pinnae bear crowded pinnules with apparently a thick lamina ;
the pinnules vary considerably in size, shape and venation, they may
be stout and broad and almost semicircular in form, deltoid or broadly
oval with slightly divergent forked veins passing into broad linear seg-
ments with a well-marked midrib giving off forked secondary veins
at an acute angle. The pinnules are confluent at the base. In the

* Ettingshausen (.52). t Morris (4-5), pi. vi.
+ Szajnocha (88). § Solms-Laubach (99), pi. xiv., fig. 2.

II Geinitz (76), pi. i., tig. 17. IF Feistmantel (77), pis. xxxix. and xlvi.

** Curran (85), pi. ix., fig. 3. ft Tenison-\Yoods (88).

54 Annals of tJic South African Mnscuvi.

shorter pinnules the lamina is traversed by several slightly diverging
veins which spring separately from the rachis, but as the pinnules-
become larger the veins tend to converge in the basal portion
of the lamina, forming a midrib (pi. vii., fig. 8). In the apical portion
of the pinnae the divisions between the segments become shallower
and the apex consists of a bluntly terminated entire lamina. Epider-
inis composed of polygonal cells or — above the veins — of oblong
rectangular cells with very slightly undulating walls. Stomata not-
very numerous, occurring on both the upper and lower epidermis,
probably slightly sunk, bordered by two or four cells which may be
subsidiary cells above the true guard-cells ; the pore of the stoma is
bounded by two crescent-shaped cuticular ridges which may belong
to the guard-cells.

The specimens from the Stormberg beds are all sterile, and none
have been found showing the habit represented in Feistmantel's
figures of Australian examples.

The most common form of pinna and pinnules is that represented
in fig. 1, pi. vii., in which the segments have no midrib but agree in
their venation with the genus Odontopteris. A few specimens occur
in which the pinnules are broadly linear and provided, at least in the
basal portion, with a midrib: these examples (pi. vii., figs. 7, 8) are, I
believe, specifically identical with those bearing shorter and broader
segments. Morris '■'- referred the pinnae with longer segments to
a variety — var. lancifolia, and by other authors this form has
been raised to specific rank.

Geinitz i figures an example of a pinna with longer pinnules
from the Argentine which he names Thinnfeldia tcmcinervis, and
Solms I refers a Chilian specimen of Eh^tic age to T. lancifolia
Morris.

Szajnocha§ also adopts a distinctive name, T. lancifolia, for
Argentine Ehaetic specimens with linear pinnules. The occurrence
of intermediate forms of pinnules connecting those having almost
parallel veins with the longer ones, in which a midrib is well defined,
leads me to include both under one specific name. A similar course
is adopted by Feistmantel || in his Australian monograph. The
Rhaetic specimens described by Nathorst from Scania and placed in
the genus Ptilozaniitcs exhibit a similar variation in the form and size
of the ultimate segments.

The Queensland specimen figured l)y Carruthers is in the British

* Morris (4-5), pi. vi., fig. 4. f Geinitz (76), pi. i., tigs. 17, 18.

+ Solms-Laubach (99), pi. xiv., tig. 2. § Szajuocha (88).

1; Feistmantel (90), pi. xxix.

Fossil Floras of Cape Colony. 55

Museum Collection (V. 4197) ; it is clearly identical with the South
African examples, but the details are less distinct in the specimen
than in the published figure.

Plato VII., fig. 1 (21e).

Branched piece of frond ; the unbranched portion shown in the
figure has a length of 11-5 cm. and is incomplete distally. The longest
segment is 1'5 cm. long and 6 mm. broad. The pinnules show very
clearly the slightly divergent and forked veins and the leathery
texture of the lamina.

Plate VII., figs. 8, 8a (45e).

Preserved as a clearly defined brown impression with numerous
fragments of pinnae with pinnules varying considerably in size ;
the figured segments show the midrib and lateral veins very clearly.

Plate VII., fig. 7 (477d).

An indistinct impression in oxide of iron on sandstone. The
linear pinnules are confluent at the base, forming a narrow wing to
the rachis. A midrib is seen in the lower part of some of the
segments, also oblique secondary veins. Associated with fragments
of pinnae bearing the shorter and broader type of segment.

Plate VIII., fig. 7 (22e).

A piece of a pinna with larger segments (1'5 cm. x 10 mm.) ; also
a forked pinna with smaller pinnules.

Plate VIIL, fig. 8 (d).

A specimen on a piece of shale from Cyphergat showing much
smaller pinnules ; associated with other specimens of the type
represented in fig. 1, pi. vii.

Plate IX., fig. 7 (A).

A few epidermal cells of a small leaflet of Thinnfeldia odontop-
tcroides, showing the slightly wavy form of the walls, and four cells
surrounding a stomatal pore. The narrower central part of the pore
is bounded by the remains of dehcate cell membranes with thickened
borders which possibly represent cuticular ridges of the guard cells.
The appearance of the stomata suggests that the guard-cells are
slightly sunk.

56

Annals of the South African MuseiLm.

Plate IX., fig. 8 (B).

A fragment showing the thickened cuticular edge of a leaflet and
a single stoma with two crescent-shaped cuticular ridges bordering
the pore.

Plate XL, fig. 2 (44e).

A forked portion of a frond showing a variation in the form of the
pinnules, some of which are identical with those represented in
fig. 1, pi. vii., while others are longer and illustrate the convergence
of the veins to form a midrib in the basal portion of the segments.

Text-figure 6 (x IJ).

Thinnfeldiu odontoptero'ules (Morr.).

From a specimen from Cyphergat in the British Museum (V. 2490).

Text-figure 6. The British Museum specimen from Cyphergat
shows the venation of the pinnules very clearly, it is undoubtedly
specifically identical with the example represented in pi. vii., fig. 1.

Other Sjjeciviens : A. Fragments from Cyphergat ; 4, 5 (associated
with a piece of Baiera). 23e, specimen with pinnules intermediate
between those shown in pi. vii., figs. 1 and 8 ; associated with
Steno2)teris elongata (Carr.) and Callipteridiwm. 24e, 25e, 26e

Fossil Floras of Cape Colony. 57

(with Taniopteris), 28e, 29e, 30e, 32e, 33e — several fragments, one
showing the apical portion of a pinna ; 35e, 36e — a good specimen
of a forked portion of a frond ; 42e, 43e, 309d, including a good
impression of the apex of a pinna, 310d, 312d, 313d, 315d, 316d,
475d, 476d, 478d, 479d.

Thinnfeldia rho:\i];oidalis Ettingshausen.

Plate VIII., fig. 1.

1852. Thinnfeldia rlwmhoidalis Ettingshausen, Abk. k.-k. geolog.

Eeichs. Wien, Bd. I., Abt. iii., p. 2, pi. i., figs. 4-7.
1867. Thinnfeldia decnrrens Schenk, Foss. Flor. der Grenzschichten,
p. 114, pi. xxvi., figs. 1-5.
T. rhomboidalis, ibid., p. 116, pi. xxvii., figs. 4, 5.

Similar in habit to Tliinnfcldia odontopteroides, but characterised
by the longer broadly linear segments decurrent by the lower margin
on the rachis, the upper edge of the base l)eing separated from the
rachis by a distinct sinus ; a well-marked midrib giving off oblique
forked veins, the lower basal portion of the segments contains a few
curved veins which enter tbe lamina direct from the rachis.

The single fragment from Stormberg represented in pi. viii., fig. 1
may belong to a larger pinna of Thiniifeldia odontopteroides , oi- it
may be included under Ettingshausen's species '•' as more clearly
resembling the Steierdorf examples than the typical forms of Morris's
species.

Schenk's figures of T. rhomboidalis f represent specimens which
■cannot be distinguished from that shown in pi. viii., fig. 1. An
example of Thinnfeldia from the Burrum Coalfield, Queensland,
referred by Jack and Etheridge I to T. media Ten. -Woods, is probably
identical with T. rhomboidalis.

Plate VIII., fig. 1 (C).

This fragment shows very clearly the form and venation of the
segments ; the agreement between this specimen and Schenk's
Thinnfeldia decurrens is exceedingly close. A fragment of Stenop-
teris elongata (Carr.) is seen resting on the Thinnfeldia.

* Ettingshausen (52), pi. i.

t Schenk (67), pi. xxvii., figs. 4, o.

+ Jack and Etheridge (92), pi. xviii., fig. 10.

58 Annals of the South African Museum.

Genus CALLIPTEEIDIUM Weiss.
Callipteeidium stokmbeegense sp. nov.
Plate VII., figs. 4-6a; Plate VIII., fig. 5.

Frond bipinnate, rachis stout ; piniiiB alternate, linear, tapering to
an acuminate apex. Pinnules linear, entire or serrate, becoming
more crowded in the distal region of the pinnse and near the apex of
the frond, confluent basally ; the rachis bears pinnules that extend
upward on to the lower side of the pinnae. Venation of the Claclo-
phlebis type, approaching the Sphcnoptcris type in the smaller
fronds (pi. viii., fig. 2).

The Ehoetic species figured by Jaeger as Aspidioides stuttcjardiensis,-''
by Nathorst f from Scania as LepidojJteris ottonis Goepp., and by
Schenk as Asplenites ottonis .j; resembles the African fern in having
pinnules on the rachis, bvit differs in the larger ultimate segments.
This European type is, however, characterised by the occurrence of
numerous scales on the rachis, a feature not exhibited by the
Stormberg plant. A French species figured by Saporta as Lomatop-
teris Hasina Morr j affords another example of a bipinnate frond
with pinnules springing from the main rachis.

In order to avoid the institution of a new generic name, I propose
to speak of the African fern as Callipteridiiim stormbergense. The
fragment of a Gladophlchis figured by Feistmantel as Asplcnium,
cf. nehbense,\\ may possibly be a piece of a large pinna of
Gallipteridiiivi storvibergense, but the specimen is too small to
determine.

One striking feature of the bipinnate frond shown in fig. 6, pi. vii.,
is the occurrence of pinnules on the main rachis ; in other respects
the characters appear to he such as might be expressed by the use of
the generic name Cladophlcbis. The Palaeozoic type Callipteris
conferta'^ bears pinnules on the rachis, and a similar character is
seen in fronds included in Weiss' genus Galliptcridium; "•"•' similarly
a fern named by Heyer to Callipteris schenkii ]]■ exhibits the same
features but bears lobed pinnules.

A fragment figured by Szajnocha || from Cacheuta (Argentine) as

* Jaeger (27), pi. viii. f Nathorst (78), pi. ii., fig. 1.

I Schenk (G7), pis. xi. and xiv. § Saporta (91), pi. Ixvii., fig. 6.

II Feistmantel (89), p. (J8, pi. ii., figs. 12, 12«.

11 See Schimper (74), Atlas, pi. xxxii., fig. 2. See also Fontaine and White (80) >
pi. xi., figs. 1-4.

•* Zeiller (00), p. 92, fig. 92. ft Heyer (84).

+ J Szajnocha (88) pi., i., fig. 8.

Fossil Floras of Cape Colony. 59

Pecoptcris sclioenleiniana Brongn. bears a resemblance to the Storm-
berg plant, but it is too small to determine.

Plate VII., fig. 6 (39e).

Eachis stout, ll'S cm. long. The linear confluent pinnules are for
the most part entire, but on the lower pinnae some show a distinct
serrate margin (fig. 6a). In the apical region the pinnules are more
crowded and gradually pass into segments with a serrate margin.
The venation is obscure, but apparently of the Cladophlehis type.
The pinnules on the lower side of the pinna3 pass on to the rachis,
where they occur as segments of varying size and shape.

Associated with fragments of Thinnfeldia.

Plate VII., fig. 4 (10).
Pinnules enlarged, showing a well-defined midrib and lateral veins.

Plate VII., fig. 5 (31e).

The apex of a frond, showing the passage of pinnules into a
terminal acuminate lamina with a sei'rate edge.

Plate VIII., fig. 2 (46e).
Habit similar to that of the larger frond shown in fig. 6, pi. vii.
The pinnge are more crowded, and the segments little more than
weU-marked serrations. Venation clearly preserved, each pinnule
receives a vein which gives off a few branches at an acute angle.
The lowest segment on the pinnae tends to be decurrent on the main
rachis.

Other Specimens : 8, small pinnse with crowded pinnules, pinnules
occur also on the main rachis ; 40e, a portion of a frond rather less
than that shown in fig. 6, pi. vi., and intermediate between that
specimen and the still smaller example represented in fig. 2, pi. viii.
480d, an indistinct impression on sandstone of part of a large frond;
associated with Thinnfeldia.

Genus T^ENIOPTERIS.

T^NiOPTERis CARRUTHERSi Tcnison-Woods.

Plate VIIL, figs. 5, 6.

1872. Tceniopteris Daintreci Carruthers, Quart. Journ. Geol. Soc,
vol. xxviii., pi. xxvii., fig. 6, p. 355.

60 Annals of the Sonth African Museum.

1883. Taniopteris Carruthersi Tenison- Woods, Proc. Linn. See.
N. S. Wales, vol. viii., p. 117.

1889. Tteniopteris Carruthersi Feistmantel, Abh. k. bohm. Ges.

Wiss., Bd. III., Folg. vii., p. 65, pi. ii., figs. 6-10.
? T. Daintrcei, ibid., p. 66, pi. ii., fig. 11.

1890. TcBniopteris Carruthersi Feistmantel, Mem. Geol. Surv.

N. S. Wales, p. 115, pi. xxviii., fig. 7.

Frond simple, strong midrib from which numerous single or forked
secondary veins are given otf either at right angles or at a wide
angle. The breadth of the lamina decreases gradually towards the
petiole.

In 1872 Carruthers '•' referred some Tceniopteris leaves from
Queensland to McCoy's species T. daintreci,] but as Tenison-
Woods pointed out in 1883, McCoy's type is characterised by the
narrow linear form of the frond and differs considerably from the
iDroader form of leaf figured by Carruthers. The specimen described
by Carruthers [in the British Museum Collection (V. 4199)] is clearly
identical with those from South Africa figured by Feistmantel. A
leaf from the Argentine named by Geinitz Taniopteris mareyiaca,X
bears a close resemblance to T. carruthersi, but there appears to be
a slight difference in the venation. Some specimens from India
referred by Feistmantel to Tanioptcris vittata,i a species differing
but sliglitly from T. carruthersi, are very similar to the type of
frond shown in pi. viii., fig. 6. A Ehastic species from Scania,
named by Nathorst Ta>,niopiteris imincrsa,\\ may be identical with the
South African fern ; but the comparison of sterile fragments of
Tceniop)teris fronds can lead to no certain results.

Plate VIII., fig. 6 (497d).

Portion of leaf 8'5 cm. long and 3 cm. broad ; the midrib is com-
paratively broad and well defined. The secondary veins are approxi-
mately 1 mm. apart or slightly closer.

Plate VIII., fig. 5 (lie).

This much narrower fragment is distinguished by the more open
arrangement of the forked and simple lateral veins, but it may

* Carruthers (72).

t For figures of the true Tainoptcris Daintrcei McCoy, see Oldham and Morris
(63), pi. vi. ; McCoy (74), pi. xiv. ; Feistmantel (77==), pi. i., &c.

+ Geinitz (76), pi. ii., figs. 1-3. § Feistmantel (76), pis. i., ii.

II Nathorst (78), pi. xix., fig. 6.

Fossil Floras of Cape Colony. 61

represent the lower and gradually tapering portion of Tanioptcris
carruthersi. Compare Feistmantel's figures of narrower fronds
referred to this species.'''

Plate VIII., fig. 4 (1).

This specimen differs from that shown in fig. G (497d) in the
greater inclination of the secondary veins and in the texture of the
lamina, which gives the impression of being woolly. Possibly it
would be wiser to distinguish this form by another specific name,
but the apparent difference in texture may be due merely to the
manner of preservation, and in an}'' case the material is too meagre
to be described under a new specific title.

Other Specimens : 2e and 3, half of a leaf lamina, similar to the
specimen shown in fig. 4, but the veins are more clearly preserved
and given off at a wider angle from the midrib ; 9e, pieces of fronds,
a broader fragment like that shown in fig. 6, and a narrower speci-
men with crowded veins ; lOe, 27e, 466d, 498d, three pieces, one
showing the gradual decrease in breadth of the lamina towards the
base of the leaf.

Genus CHIEOPTEEIS Kurr.

This genus, instituted by Kurr for a leaf from Keuper beds near
Stuttgart, is thus defined by Schimper : — j

" Frons petiolata, simplex, irregulariter inciso-digitata, coriacea.
Nei'vi numerosissimi e basi radiantes, tenues, omnes, asquales vel
loborum nervo medio distinctiore, repetito-dichotomi, anastomosantes,
reteque anguste rhomboiderm efficientes."

We possess no satisfactory evidence as to the afiinity of this
genus. In form and venation Chiropteris agrees with the recent genus
Opliioglossum, and on the strength of this resemblance it has been
placed by some authors in the Ophioglossaceae.|

Zeiller j has also referred to the resemblance between Chiropteris
and Ophiofjlossiou, but a comparison based on form and venation
only is in this case of little or no botanical value. The genus may
at present be included provisionally among the ferns, with the
reservation that we lack satisfactory evidence as to its taxononiic
position. The leaves from the Kootanie group referred Jjy Newberry

* Feistmantel (89), fig. 7. t Schimper (69), p. 643 ; Atlas, pi. xliii.

+ Schimper and Schenk (90), p. 152 ; Potonie (99), p. 91.
§ Zeiller (00), p. 55.

62 Annals of the South African Museum.

to CJiiropteris should probcably be included in a distinct genus, one
of them — C. spatulata — may be a species of Sageii ptaris:''' The
Chilian fossil referred by Solms-Laubach | to Kurr's genus differs in
the form of the leaf and in the more crowded veins from the African
species.

Chieopteris cuneata (Carruthers).
Plate IX., fig. 4.

1872. Cycloptcris cuneata Carruthers, Quart. Journ. Geol. Soc,

vol. xxviii., p, 355, pi. xxvii., fig. 5.
1889. ? Anthrophiiopsis sp. Feistmantel, Abh. k. b5hm. Ges. Wiss.

Folge, vii., Bd. III., p. 67, pi. ii., fig. 4.

Leaves cuneate in form, the lamina traversed by numerous
spreading veins, dichotomously forked and occasionally anastomosing
by oblique cross veins.

Carruthers thus defined the species Cyclopteris cuneata, which he
founded on a specimen from the Tivoli Coal-mine, Queensland : —

" Form of the entire frond unknown ; pinnae entire, large, cuneate,
narrowed at the base, with the distal margins rounded ; veins deli-
cate, once or twice dichotomously divided ; sometimes anastomosing
once in their length near the middle of the pinna." j

The type-specimen in the British Museum (V. 4197) is accurately
represented in Carruthers' figure ; the veins are very obscure. The
anastomosing veins which are described by Carruthers and shown
in the specimen represented in pi. ix., fig. 4, and still more clearly
in a better specimen from the Stormberg beds in the British Museum
Collection (V. 2498), render Cycloptcris unsuitable as a generic
designation.

It is by no means unlikely that the fragment of a leaf figured by
Feistmantel from Cyphergat d.s Anthrophyopsis (?) § may be identical
with the plant represented in pi. ix., fig. 4. An imperfect specimen
figured by Fontaine from North American Triassic beds || as Sagen-
opteris rhoifolia bears a distinct resemblance to the Stormberg leaf.

Plate IX., fig. 4 (C).
The distal margin of the wedge-shaped leaf is rather torn and
incomplete ; the thin lamina tapers towards the base, which is

* Newberry (91), p. 198, pi. xiv., figs. 1, 2, 10, 11.

f Solms-Laubach (99), pi. xiii., figs. 1-4.

\ Carruthers (72), p. 355, pi. xxvii., fig. 5.

% Feistmantel (89), p. (J7, pi. ii., fig. 4. || Fontaine (83), pi. xlix., fig. 5.

Fossil Floras of Cape Colony.

63

also only partially preserved. The veins are fairly distinct ; they
diverge from the base, forking and occasionally anastomosing by
oblique cross-connections.

Associated with Thinnfeldia and Plicpuicopsis.

Other Specimens : h. An impression similar to that figured (pi. IX.,
fig. 4) ; veins fairly clear. Associated with TkhmfeUlia and
Pho&nicopsis.

Chiropteris zeilleri sp. nov.

Text-figure 7.

Leaf petiolate ; lamina more or less orbicular, deeply lobed into
broad segments with an irregularly lobed margin ; veins spreading
from the base, forked and anastomosing, forming long and compara-
tively broad polygonal meshes.

Text-figure 7. (Nat. size.)

Chiropteris zeilleri sp. nov. From a specimen from Cyphergat in the British

Musemn (V. 3268).

This species is founded on a single specimen from Cyphergat in
the British Museum Collection (V. 3268), shown in text-figure 7.
The t}^De- specimen is an imperfect leaf showing the upper part of a
jDetiole bearing a deeply lobed lamina, 7-5 cm. in depth and S-5 cm.
broad ; the sinus on the right-hand side of the lamina is preserved
intact, while those on the left are partially torn.

64 Annals of tlie South African Museum.

The veins, which are very clearly preserved, radiate from the apex
of the leaf-stalk, and by the occasional anastomosis of adjacent
forked branches, form long meshes.

In shape the leaf may be compared with small fronds of the
recent fern Dipteris,''- and with the fossil genera Protorhipis of
Andrae and Hausmannia of Dunker, but the absence of veins in the
meshes constitutes an important distinguishing feature. Some
Jurassic Bornholm ferns figured by Moller in a recent paper and
referred to Haiisviannia,]- but which I should prefer to assign to
Frotorhipis, illustrate the superficial resemblance of the Dipteris
type of frond to Ghiropteris zeilleri.

A fossil leaf with somewhat similar, but not identical, venation
has been figured by Feistmantel from India as Belcmnopteris Wood-
masoniana X and compared with the recent genus Hemionitis. The
genus Hemionitis,^ a typical American member of the Polypodiacese,
is characterised by its palmate lamina and prominent ribs, thus
differing from Ghiropteris. The recent genera Opliioglossimi and
Antrophyum afford the nearest parallels as regards venation to the
fossil frond, but we have absolutely no evidence as to the pi'ecise
position of Ghiropteris.

The specimens of Gliiropteris figured by Bronn from the Letten-
kohle beds jj differ from G. zeilleri in the much more numerous and
crowded veins, which are not described as anastomosing, and in the
apparently whorled disposition of the sessile leaves. I have named
the Stormberg leaf after my friend Professor Zeiller of Paris, whose
work has done so much to extend our knowledge of the vegetation
of the Southern hemisphere.

GYMNOSPEEMiE.
Group GINKGOALES.

Genus BAIERA, Braun.

Baieea stormbekgensis sp, nov.

Plate VIII., fig. 3.

Lamina deeply dissected into broad linear segments which are
again sulxlivided distally into narrower linear lobes ; veins com-

* Seward and Dale (01). t Moller (02). + Feistmantel (80), pi. xliii. A.

§ Christ. (97), p. 57. || Bronn (58), p. 143, pi. xii.

Fossil Floras of Cape Colony. 65

paratively few in the lower pai-t of the lamina, but after repeated
dichotomy they become more numerous in the upper part of the
leaf, where they occur as crowded parallel veins slightly converging
towards the tips of the narrow terminal segments. The whole
lamina reached a length of 12 cm. or more and probably a breadth
of 10 cm.

I have given a new specific name to the unusually large type of
leaf represented in pi. viii., fig. 3 ; it might with equal propi'iety
be included in the genus Ginkgo, as one occasionally finds large and
deeply lobed leaves of the Maidenhair tree very similar to the example
shown in the figure."

The species described by Feistmantel from the Stormberg beds as
Baiera schenki f may be merely a smaller form of B. stormbergensis,
but the much greater breadth of the segments and the more open
character of the venation constitute distinctive features. It is,
however, by no means improbable that the narrower and broader
forms may be found to be inseparable. Feistmantel's B. schencki
bears a fairly close resemblance to B. longifolia Heer.]:

Other large leaves comparable with B. stormbergensis have been
•described by Shirley from Ipswich, Queensland as Ginkgo simmondsi,
Baiera ginkgoides and B. ipsviciensis ; j Eatte || has also described
a very large leaf from the Wianamatta shales under the name
Salisburia palmata. It is possible that the large Australian leaves
referred by Shirley to Ginkgo sirmnondsi and by Ratte to Salisburia
palinata may be identical with the Triassic American type Baiera
multifida.*',

Plate VIII., fig. 3 (x).

The lamina is incomplete, but reaches a length of 11 cm. The
spreading and repeatedly forked veins are very clearly shown
traversing the long lobes of the leaf, which is preserved in the
form of a light brown impression, which suggests a thin lamina
similar to that of Ginkgo biloba.

Other Specimen : 16c. A small specimen showing terminal por-
tions of a leaf and slightly converging veins at the tips of the
segments.

* Seward and Gowan (00), ]i\. x., fig. 63.
t Feistmantel (89), p. 72, pi. iii., figs. 1, 2, 5, 6.

: Heei- (77), pis. vii.-x., xv. § Shirley (98). || Eatte (88) (88').

•[ Fontaine (83), pis. xlv.-xlvii. See also Seward and Gowan (00), p. 139, and
Arber (02), pp. 4, 5.

66 Annals of the South African Museum.

? GINKGOALES.
Genus STACHYOPITYS Schenk.

Stachyopitys sp.

(Cf. Sjjhenoleins rhatica Geinitz).

Plate IX., figs. 2, 2a.

A fragment of a fertile shoot consisting of a fairly broad axis-
bearing lateral branches terminating in a cluster of obovate bodies
arranged in apparent whorls.

The nature of the fossil represented in pi. ix., figs. 2, ^a, must be
left in doubt ; the single specimen is too incomplete to afford
sufficient evidence as to its structure and affinity. A somewhat
similar fragment is figured by Geinitz '■' from the Ehaetic of San
Juan (Argentine) as Sj)henolepis rhcBtica ; he describes the specimen
as consisting of spherical or compressed cones bearing numerous
spirally disposed woody scales of obovate form, their outer convex
surface exhibiting indistinct and irregular longitudmal striae or
grooves. In size, and in the appearance of the cones, Geinitz's
specimen agrees with that shown in fig. 2 ; but in the African
example the scales occur in a single radiate series with portions of
a superposed second series seen at a, fig. 2a.

Another fossil bearing a resemblance to the Stormberg specimen
is that described by Schenk under the name Stachyointys preslii f
as male flowers of a Conifer from the Ehtetic of Franconia ; the chief
difference between the German and African fossils is the smaller
number of radiate appendages in the former. A Queensland fossil
described by Shirley as Stachyopitys annularioides \ also agrees
fairly closely with the Stormberg fragment, so far as it is possible
to base an opinion on the somewhat crude drawings.

From Thuringian Permian rocks Potoni6 has figured fertile
portions of Pecopteris lyinnatifida (Gutb.) § which present an appear-
ance similar to that of Schenk's Stachyopitys preslii and to the
specimen shown in pi. ix., fig. 2.

Without expressing a definite opinion as to the botanical position
of the fossil, it may be provisionally included in Schenk's genus
Stachyopitys, with the suggestion that it may be a fragment of a

* Geinitz (76), p. 12, pi. ii., figs. 23, 24.

t ScJienlc (G7), p. 185, pi. xliv., figs. 9-12.

+ Shirley (98), p. 13, pi. xviii.

§ Potonie (93), pi. xviii., figs. 9, 10, and pi. x.

Fossil Floras of Cape Colony. 67

male flower of a uioiiher of the Ginkgoales. The radially disposed
oval bodies may be microsporangia differing chiefly in their greater
number from the usual type of the fossil male flowers of the Gink-
goales.

Plate IX., figs. 2, 2a (12).

A fragment of an axis 1-5 mm. broad with the remains of swollen
lateral branches bearing circular rosettes 5 mm. in diameter consisting
of fairly stout oval bodies attached ))y a narrow base and terminat-
ing in a bluntly rounded apex ; these oval bodies are thick and
rounded, and their surface shows a few longitudinal ridges. At a
(fig. 2a) one sees the remains of other similar appendages superposed
on the lower and more perfectly preserved ones.

PLANTS OF DOUBTFUL POSITION.
Genus PHCENICOPSIS Heer.

In instituting this genus for Jurassic specimens of linear leaves
from Siberia, Heer defined it as follows : —

" Folia coriacea, numerosa, inramulo abbreviato caduco fasciculata,
squamis compluribus persistentibus cincta, sessilia vel in petiolum
brevem sensim attenuata, indivisia, multinervia, nervis simplicibus,
parallelis, densis." -•'

We have no satisfactory evidence as to the botanical position of
Phoenicopsis, but it has been referred to by Heer and other authors as
closely allied to Baicra and possibly a member of the Ginkgoales.

Phcexicopsis elongatus (Morris).
Plate IX., figs. 1, 9, 10.

1845. ZciKjopliyllitcs elongatus Morris, in Strzelecki's New South

Wales, p. 250, pi. vi., figs. 5, 6a.
1882. Cf. Nteijgerathiopsis Hislopi Zeiller, Ann. des Mines,

vol. li. [8], pi. xii., fig. 11.
1889. Podozamites {Zeugophyllites) elongatus Feistmantel, Abh.

K. bohm. Ges. Wiss., Bd. III., Folg. vii.,p. 68, pi. ii.,

fig. 13; pi. iii., figs. 3, 4, 7.

Leaves linear, reaching a length of more than 16 cm. ; the lamina
is usually gradually tapered to an acuminate termination which may

* Heer (77), p. 49.

€8 Annals of the South African Museum.

represent the basal end, but occasionally the leaf ends in a bluntly-
rounded apex which may be the distal termination.* Veins numerous,
parallel, usually simple and not dichotomously branched. The
numerous parallel veined linear leaves from Cyphergat, of which
portions are shown in pi. ix., figs. 1 and 9, agree closely with
Morris's species, Zeugophyllites elongatus, from the Jerusalem Basin,
'Tasmania.! The type-specimen of Morris's Jerusalem Basin plant,
Zeugophyllites elongatus, which has recently been found by Mr. Arber
in the British Museum Collection (V. 7512), convinces me of the
identity of the Tasmanian and South African leaves.

The same name has been applied by Szajnocha| to a fragment
from Cacheuta, which may not, however, be identical with Morris's
type. Some specimens of linear leaves from South America in the
British Museum Collection are undoubtedly identical with Morris's
species. Feistmantel has figured several leaf fragments from the
Stormberg beds which he speaks of as Podozamites {Z eugophyllites)
elongatus (Morris) ;§ these are no doubt identical with the specimens
represented in pi. ix., figs. 1, 9 and 10. It is possible that these leaves
maybe correctly referred to the genus Podozamites, but I am disposed
to regard them as more probably examples of Heer's genus PJioinicopsis,
similar to P. spcciosa of Jurassic age.|| Krasser has figured several
leaves very similar to mine from China which he refers to a new
species Phcenicopsis media IF ; one of his figures shows clearly that
the leaves converge in groups to a common point of origin, but in
the scattered fragments from Cyphergat it is impossible to recognise
any such orderly arrangement. It is very difiicult to determine
with any degree of accuracy isolated leaves of the type of
Phcenicopsis or No&ggerathiopsis ; Zeiller's Tongking specimens may,
I believe, be identical with Morris's species from New South Wales,
but as Prof. Zeiller points out, in a letter written in reply to my
suggestion, there is no evidence in the specimens he has seen to
warrant the conclusion that they were borne in bundles like Heer's
examples of Phcenicopsis. k similar leaf is figured also by Broun
from the Lettenkoble of Eaibl ** as a monocotyledonous leaf. The
bluntly terminated leaf fragment from Molteno shown in fig. 10

* Cf. Heer (77), pis. xxviii., xxix.

t Morris, in Strzelecki (45), p. 250, pi. vi., figs. 5, 5rt.

\ Szajnocha (88), pi. ii., fig. 4.

§ Feistmantel (89), p. 68, pi. ii., fig. 13 ; pi. iii., figs. 3, 4, 7.

11 Heer (77), pis. xxviii., xxix.

Ii Krasser (00), pi. iii., fig. 4; pi. iv., fig. 2.

** Bronn (58), p. 135, pi. vii., fig. 4.

Fossil Floras of Cape Colony. 69

suggests a comparison witli the genus Podozamites, but this occurs
in association with gradually tapered leaves like that I'epresented
in fig. 9, and the difference in the method of termination is in
itself hardly a sufficient reason for the separation of the two
forms.

With the exception of the forked vein shown in fig. 10, pi. ix., near
the apex of the specimen, I have not detected any certain indication
of the parallel veins in these leaf fragments from Cyphergat and
Molteno ; in Nrnggcrathiopsis leaves of similar form to that shown in
pi. ix. figured by Feistmantel,''' Zeiller,! and other authors the veins
are dichotomously branched in their course through the lamina, and I
have no hesitation in expressing the opinion that the African specimens
should not be referred to that genus. Kurtz \ has recently expressed
a similar opinion as to the difference lietween the true Noaggerathiopsis
and Morris's Zcugophyllites clongatus.

Mr. x\rber § has also drawn attention to a confusion between Noegger-
athiopsis and Zcugoplvjllites.

Plate IX., fig. 1 (C).

An imperfect leaf 16 cm. long, and with an approximately uniform
breadth of 1'2 cm. The lamina between the fairly broad veins is
finely striated, the stri» no doubt representing hypodermal strands
of mechanical tissue.

Plate IX., fig. 9 (d).

The figure represents the gradually tapered (basal ?) end of an
incomplete leaf 10.5 cm. long.

Plate IX., fig 10 (69).

The blunt apex of a leaf from Molteno showing the converging
terminations of the veins. The venation is less clearly seen
in this specimen than in the other figured pieces. Acuminately
pointed leaves similar to the specimen from Cyphergat represented
in fig. 9 occur on the same piece of rock.

* Feistmantel (80), pi. xlv. A. t Zeiller (82), pi. xii., fig. 11.

+ Kurtz (03), p. 25. § Arber (02), p. 18.

70 Annals of the South African Museum.

Genus STENOPTEEIS Saporta.

Stenopteris elongata (CaiTuthers).

Plate VII., figs. 2, 3.

Plate XI., fig. 3.

1872. Sjjhemptci-is elongata Carruthers, Quart. Journ. Geol. Soc.

vol. xxviii., p. 355, pi. xxvii., fig. 1.
1883. Sphcnopteris elongata Tenison-Woods, Proc. Linn. Soc. N.
S. Wales, vol. viii., p. 92.
Trichomanitcs sinnifolmm, ibid. p. 95, pi. iii., fig. 7.

1888. Si)hcnoi)teris elongata, Szajnocha, Sitzb. Kais. Akad. Wiss.

Wien., Bd. XCVII., Abth. i., p. 5., pi. ii., fig. 2.

1889. S-plienopteris elongata Feistmantel, Abh. k. bohm. Gess.

Wiss., Bd. III., Folg. vii., p. 61.

1890. Sphenopteris elongata Feistmantel, Mem. Geol. Surv. N. S.

Wales, p. 90.
Trichomanites spinifolium, ibid. p. 95.
1892. Trichomanites spinifolium Jack and Etheridge, Geol. Pal.

Queensland, p. 315, pi. xviii., fig. 8.
1898. Trichomanites elongata var. spinifolia, Shirley, Bull. no. 7,

Geol. Surv, Queensland, p. 19, pi. x., fig. 3.

Fragments of a repeatedly branched narrow woody axis, the ulti-
mate branches alternate, linear or short and pointed. The linear
branches may show what appears to be a comparatively broad midrib
(pi. xi., fig. 3), but usually there is no clear indication of a median
vein. In 1872 Carruthers described a plant from the Tivoli Coal-
mine, Queensland, as Sphenopteris elongata, which he defined as
follows : —

" Frond dichotomously divided, each division irregularly pinnate ;
pinnae simple, liifurcate, or irregularly pinnate ; segments narrow
linear, slightly tapering upwards to the somewhat blunt apex ;
the single midrib sending out simple branches, which run along
the middle of each segment." "

Carruthers mentions the occurrence of sori as small oval markings
scattered irregularly on either side of the midrib, but no trace of such
markings are shown in the published figure. In some of the frag-
ments of this plant in the British Museum Collection the central rib,
which stands out clearly in Carruthers' figure, is much less distinct

* Carruthers (72), p. 355, pi. xxvii., tig. 1.

Fossil Floras of Capa Colony. 71

and the winged membranous border represented in the drawing is not
shown in the specimens that I have examined. A few of the Tivoli
specimens in the British Museum are characterised by more
numerous lateral branches and form connecting links between
the types represented in pi. vii., figs, la and 3.

Szajnocha '■'■ refers a small specimen from Argentine Ehtetic beds
to Carruthers' species, but his material is too meagre to afford addi-
tional evidence as to the nature of the plant.

There can be no doubt as to the identity of the specimens shown
in pi. vii., figs. 2, 3, and pi. xi., fig. 3, with Carruthers' species, but I
have no hesitation in describing the South African examples as too
stiff and woody to be referred to Splienopteris. There is no well-
defined median rib in the majority of the specimens, but the striated
appearance of the impression and the stout form of the branches
point to a plant of more woody texture than the frond of a fern.

In all probability the specimen represented in pi. vii., fig. 3, is
specifically identical with those represented in fig. 2. The smaller
form with the pointed and shorter branchlets agrees very closely with
some Australian fossils figured by Tenison- Woods f as Trichomanites
sjnnifolium but afterwards referred by Shirley to Trichomanites elon-
gata.l Feistmantel § records a specimen of Carruthers' species
from the Hawkesbury beds of New South Wales, and adds that its
occurrence in the African Stormberg beds might be expected seeing
that it is a common plant in association with Thiniifeldia in Queens-
land and that it occurs with this genus also in South America.

It is difficult to decide what generic designation to apply to the
specimens represented in figs. 2 and 3, pi. vii., and fig. 3, pi. xi.
The genus Sphcnoptcris used by Carruthers is, I believe, inappro-
priate, as the nature of the fossils does not appear to me to favour
their identification as ferns. The lateral branches seem to be
identical in texture with the axis from which they spring, and in the
smaller specimen (pi. vii., fig. 3) the small ultimate branchlets show no
trace of a median vein but present a longitudinally striated appearance
(fig. 3«) suggesting a woody texture. I am unable to recognise any
indication of scale-leaves, nor do there appear to be any transverse
divisions or constrictions. Some specimens figured by Saporta from
the Kimeridgien of France and placed by him in his genus Stenop-
teris agree closely with the African plant in habit, but differ appa-
rently in the presence of a more distinct median vein and in the

* Szajnocha (88), pi. 11, fig. 2. f Tenison-Woods (83), pi. iii., lig. 7.

+ Shirley (98), pi. x., fig. 3, p. 19. § Feistmantel (89), p. 61.

72 Annals of the SontJi African Museum.

membranous border of the linear segments. On the other hand,.
Saporta describes the fossils as having an " evidently coriaceous
texture," '^' and he suggests the probability of Stenopteris being more
closely allied to the cycads than to the ferns. "We may provisionally
make use of Saporta's generic name, noting the absence of a well-
defined median vein, a character included in the diagnosis of the
genus as represented by the type StenojJteris desviomera of Kimeridge
age.

Some figures published by Feistmantel under the name Palissya I
bear a resemblance to Stenopteris elongata, but the Indian and African
plants are probably not identical. A comparison may also be sug-
gested between the specimen shown in fig. 3, pi. vii., and Feistmantel's
figure of Pachyphyllum (Cnjptomerites) divaricatwn,l but there is-
no evidence of the individuality of the ultimate branches in the
African specimens suggesting their identity with the leaves of
Cryp)tomerites.

Such specimens as that shown in fig. 2, pi. vii., might be expected
to show indications of nodal regions. The appearance of the fragments
recalls that of a xerophytic shrubby plant with branch-like leaves or
with branches which played the part of leaves, as in the recent genus
Psilotu7H.

A plant of somewhat similar form, but not identical, has been
described from English Wealden strata as Becklesia anomala. §

Another genus comparable with Carruthers' Sphenopteris elongata
is Schenk's Frcnelopsis, but the well-marked nodal regions in that
genus afford a distinguishing feature.

Plate VII., fig. 2 (21e).

A branched axis 11 cm. long, of woody texture. I can detect no
distinct trace of a single median rib, but here and there a slight
indication of a central ridge or depression may be seen. No trace of
leaves or of nodal lines or constrictions. None of the l^ranches
are complete.

Plate VII., figs. 3 and 3a (B.)

A graphitic impression on a dark slaty rock. The axis of the^
lateral branches is represented by a fairly distinct median rib, but in
the ultimate segments, which are for the most part entire but in
a few instances lobed, there is no median vein. Fig. 3a shows an

* Saporta (73), p. 2i)0, pi. xxxii. f Feistmantel (70), pis. xi. and xii.

+ Feistmantel (7(1), pi. x. fig. 1. § Seward (Drj), p. 17'.), pi. xiv.

Fossil Floras of Ca^K Colony. 73

ultimate segment slightly enlarged with numerous irregular longi-
tudinal striations.

Other Specimens : lie, a fragment showing the termination of a
lateral l^ranch ; 38e

Plate XI., fig. 3 (6).

A fairly well-preserved impression of an axis bearing sub-
opposite branches ; both the main axis and the lateral members-
show a comparatively broad dark band which probably indicates
the presence of a midrib, a feature not seen in the other
specimens.

III.— CONCLUSION.

The following species have been recorded as representatives of
the Stormberg Flora : —

Equisetales.

Schizoneura krci'^seri sp. nov S.W. of Matatiele, Griqualand E., Tina

River.
Strohilites sp Tina Eiver.

FiLICALES.

Thinnfcldia odontopteroides (Moit.) .... Kenigha Eiver, W. of Maclear, Griqua-
land E., S.W. Matatiele.

Thinnfeldia rhomhoidaliA Ett Stormberg.

Cladoplilehis sp. (Feistmantel) Indwe.

Callipteridiam stormher(jensc sp. nov. . . Kenigha Eiver.

Tceniopteris carrutherd Ten. -Woods .... Kenigha Eiver.

Chiropteris cuneata (Carr.) Cyphergat.

Chiropteris zeilleri sp. nov Cyphergat.

GiNKGOALES.

Baiera stormherrjemis sp. nov Kenigha Eiver.

Baiera schencki Feist Indwe, Stormberg.

Plants of Doubtful Position.

Phoenicopsis elongatns (Morr.) Indwe, Molteno, Stormberg, Cyphergat,

Steiiopteris elongata (Carr.) Kenigha Eiver.

The flora of the Stormberg series may be assigned to the Ehaetic
period, a period of which the vegetation has left remnants in nearly
all parts of the world. In a paper on Ehaetic plants from Honduras,
Prof. Newberry ='= drew attention to the uniformity of the Ehaetic
vegetation both in the old and new worlds, and added : " We shall

* Newberry (88).

74 Anncds of the South African Museum.

look now with eagerness to South America for the identification
there of this Mesozoic flora, which we have found in full development
in Virginia, New Mexico, Sonora, and now in Honduras. It has
been recognised in Austraha, New Zealand, India, Tonquin, China,
Turkestan, and various parts of Europe. Hence with its discovery
in South x\merica we shall see it reaching as a girdle around the
entire globe." Before this was written Geinitz had recorded Ehsetic
plants from the Argentine, and more recently Szajnocha, Solms-
Laubach, and Kurtz have made further contributions towards the
■completion of this girdle of Ehffitic floras.

The facts of greatest interests as regards the vegetation of this
period are— the striking departure, as regards the general facies of
the floras, from the type represented by the Palaeozoic floras, and,
secondly, its uniform character in almost all parts of the w^orld.

Schdzoneura krasseri sp. nov. — This type of Equisetaceous plant
closely resembles forms recorded from Scania, Tongking, Persia, and
India, but the accurate determination of fragments of vegetative stem-
casts is an impossible task without the evidence of reproductive
organs and the study of anatomical characters.

Thinnfeldia. — This generic type — very closely allied to Ptilozamites
and other genera — is one of the most characteristic and widely spread
Ehsetic plants. It occurs in Europe (Scania, Germany, Italy, Poland,
and elsewhere), Australia, South America, and India. In North
America a very similar type of plant is recorded from rocks assigned
to a Permian horizon.

Gallipteridium stormhergensc. — This type does not afford much
assistance in solving the question of geological age ; ferns of similar
external form occur both in Permian and Liassic strata.

TcBiiioptcris carruthcrsi. — The genus Tcsiiiopteris is so abundantly
represented in strata ranging from Triassic to Wealden in various
parts of the world that we cannot with safety base conclusions on
fragments of sterile fronds, but so far as the evidence goes it favours
the identification of the African form with Ehtetic species from other
regions.

Cliiroptcris. — This appears to be a comparatively rare type, but
such evidence as it atfords as to geological age points to a Ehtetic
horizon.

Baiera stormhcrgensis. — This unusual form of leaf appears to be a
distinct species, but such Ginkgoales species as resemble it most
closely are recorded from Australian strata that may probably be
regarded as Ehaetic.

Utachypitijs. — This fragmentary fossil cannot be determined with

Fossil Floras of Cape Colony. 75

confidence, but it is in this case also Ehaetic types with which the
Stormberg plant exhibits the greatest resemblance.

Fluenicopsis. — A widely-spread form of leaf of which the systematic
position is uncertain. Leaves from Europe, China, Australia, and
Tongking (referred by Zeiller to Naggeratliiopsis) of Ehaetic age and
species described by Heer from Arctic Jurassic rocks exhibit a very
close resemblance to the African specimens.

Steiwpteris elongata. — A little-known type occurring in South
America, Australia, and South Africa in beds which are probably of
Ehajtic age.

[Note. — Since the above was written I have received, through
the kindness of Professor Zeiller, a copy of his atlas of the " Flore
fossile des gites de Charbon du Tonkin" (Paris, 1902). Some of
the plants shown in the unusually good photographic reproductions
bear a close resemblance to Stormberg species. Schizoneiira carrerei
Zeill. is probably identical with the stem which I have named
S. krasseri ; Ctenoptcris sarreni Zeill. probably represents a type
closely allied to, but not identical with, Tliinnfcldia ; the leaves
referred to Nceggcrathiopsis liislopi may, I think, be identical with
Phoenicopsis elongata.]

76 Annals of the South African Museum.

lll.—EGCA FLOBA.

Plates X.-XIII.
Text-figure 8.
I.— INTEODUCTION.

The base of the Lower Karoo formation has been described as con-
sisting in certain regions of a conglomerate to which Dunn's term,
Dwyka conglomerate, has frequently been applied. The history of
our knowledge of this member of the Karoo formation has recently
been dealt with in the Annual Eeport of the Geological Commission
of Cape Colony for 1900.* In recent years attention has been given
to the occurrence of conglomerates in the Transvaal which are cor-
related with the Dwyka conglomerate of Cape Colony and Natal.
In 1895 Schmeisser recorded Schizoneura and Glossopteris from
Transvaal collieries in beds resting on the conglomerates, and con-
cluded that the plant-bearing strata should be correlated with Ecca
shales of Cape Colony. Dr. Corstorphine in his Eeport of 1900 t
summarises the observations bearing on the nature and age of the
conglomerates of Cape Colony and the Transvaal as follows: " The
evidence of the glacial action in the Dwyka conglomerates has
accumulated to such an extent during recent years that doubt can
no longer remain that the rock owes its peculiarities to such action,
and that the Transvaal conglomerates, associated with the coal, are
also unmistakably of glacial origin, and, from the fossil evidence
found in the overlying rocks, can equally surely be regarded as being
at the same geological horizon as the Southern Dwyka."

In 1896 Professor Zeiller X published an important paper in the
Bulletin of the Geological Society of France dealing with fossil
plants obtained from the neighbourhood of Johannesburg. These
included Glossopteris (leaves and rhizome), a species of Sijlienopteris ,
Phyllotheca, Naggerathiopsis hislojn (Bunb.), and some seeds. In
the following year Mr. David Draper ^ read a paper before the
Geological Society of London, in which he referred to certain plants
obtained from a quarry 2 miles east of Vereeniging, a township
30 miles south of Johannesburg. He expressed the opinion that

* Corstorphine (00). f Corstorphine (00), p. 17.

J Zeiller (1(0). § Draper (97).

Fossil Floras of Cape Colony. 77

the plant-l)earing strata should he assigned to a Triassic horizon.
Through the courtesy of Mr. Draper, I had the opportunity of
describing the Vereeniging plants,* which were referred to the
following genera : Glossoptcris, Gangamopteris, Nosggerathiopsis, a.nd
Sigillaria — types unmistakably indicating a Permo-Carboniferous age.
In 1898 Dr. Hatch j gave an account of the Geology of certain dis-
tricts in the Southern Transvaal, and in an appendix to his paper I
added a short description of a few plant remains from the Vereeniging
Sandstone.]: A Geological map of the Vereeniging district is pub-
lished in Dr. Hatch's memoir, and sections are also given by Mr.
Draper in his paper of 1897.

The plants from Vereeniging described in the following pages
were obtained by Mr. Leslie from the bed of the Vaal Eiver, about
3 miles from the locality where Mr. Draper's specimens were
collected. An expression of thanks is due to Mr. Leslie for the
important additions to the Pakeozoic Flora of South Africa which
we owe to his enthusiasm and ability.

In addition to these Lower Karoo plants from Vereeniging the
collection recently submitted to me includes a few very imperfect
fragments of leaves from Worcester and Tuin Kraal, a few miles
south of Prince Albert Eoad Station, obtained from beds assigned to
the Ecca series. The Ecca beds occur mainly in the Karoo, but
there is a considerable outcrop also in the Breede River Valley,
which has recently been examined by Mr. Schwarz,§ who mentions
the occurrence of Gangamopteris impressions in a quarry near
Worcester Station. As a matter of convenience the Worcester
fossils and those from Vereeniging may be dealt with separately.

4.— ECCA PLANTS FROM WORCESTER.

Feistmantel's list of plants from the Ecca series includes the
following species : — ||

Glossopteris hroioniana Brongn.

Gangamopteris cyclop)teroides var. attemiata Feist.

Nceggerathiopsis kislopi (Bunb.).

The material in my hands is too meagre and imperfect to lead to
any definite conclusion as to the age of the rocks ; it consists of

* Seward (97). t Hatch (98). + Seward (98).

§ Schwarz (98), p. 24. || Feistmantel (89).

78 Annals of the Soutli African Museum.

small and fragmentary impressions of Glossopteris fronds, with one
or two imperfect specimens of GMigamojJteris, preserved in a mottled
purple and greyish-green argillaceous rock, from Worcester, also some
indistinct portions of an Equisetaceous plant. In addition to these,
the rock from Tuin Kraal Eiver contains fragments of Equisetaceous-
stems, probably referable to the genus Schizoneura.

II.— DESCEIPTION OF SPECIMENS.

Genus GLOSSOPTEEIS Brongniart.

GLossorTEiiis sp.

(Probably Glossoptcris broiniiana Brongn. var. indica.)

Plate XIIL, fig. 1.

The fragments from Worcester are too imperfect to determine
with certainty. While they exhibit a fairly close agreement with
Feistmantel's figure of a Beaufort specimen, Glossojjtcris damudica
var. stenonoeura,-- they agree even more nearly with the large Glossop-
teris fronds, of which Zeiller has recently published some excellent
photographs under the name Glossopteris indica Schimp.f This
species Zeiller considers should not be regarded as identical with
Glossoptcris hroioniana. The African material is too meagre to
throw any fresh light on this difficult question of specific limitation,
but I have elsewhere adopted the plan of maintaining Brongniart's
name, Glossopteris hroioniana var. indica, in preference to Schimper's-
specific designation, G. indica. \. On the whole, I consider the
Worcester specimens identical with the type of frond figured by
Zeiller as G. indica, and by myself as G. hroioniana var. indica.

Plate XIIL, fig. 1 (Z).
A small piece of a leaf showing clearly defined secondary veins.

Other specimens : G. A piece of a frond at least 6 cm. broad,,
showing a strong midrib giving off numerous oblique anastomosing
veins ; X, 212a, 213a, 215a ; 220a, part of a large frond showing the.
midrib and anastomosing secondary veins ; 221a.

* Feistmantel (89), pi. iv., fig. 7. f Zeiller (02) ; Zeiller (9G), pi. ,307.

+ Seward (97).

Fossil Floras of Cape Colony. 7J>

Genus GANGAMOPTERIS McCoy.

Gangamopteris cyclopteroides (?) Feist.

One small and very obscure fragment (specimen Y) from Wor-
cester is no doubt a fragment of Gangamopteris ; it agrees closely
with Feistmantel's figure of G. cyclopteroides var. attennata from the
Kimberley series," but the specimen is too poor to determine.

Genus SCHIZONEUEA Schimper and Mougeot.

SCHIZONEURA sp. (?)

A single specimen (338) from Worcester consists of small and
very obscure impressions of what are probably pieces of a ScJiizo-
iieura stem, also a fragment of an Equisetaceous stem (351) on a
piece of sandstone from a kopje 5 miles north of Worcester.

Three other specimens (104a, 110a, Ilia) from the Tuin Kraal
River agree in the crowded internodal ribs with the Stormberg
examples shown in figs. 5, 6, pi. ix., but they may also be closely
matched with specimens of Scliizoncura cjondioanesis Feist, from
India, and with impressions from more than one geological horizon.

The available evidence is insufficient to throw much light on the
age of the W^orcester and Tuin Kraal rocks, but it is not unfavour-
able to the view that the beds in question are correctly placed on the
horizon of the Ecca series.

5.— ECGA PLANTS FROM VEREENIGING. f

The specimens recently collected in the Vereeniging sandstones
by Mr. Leslie are of especial interest as including three generic
types, Psygmophyllian, Neuropteridiuvi, and a Lepidodendroid stem
hitherto unknown from South Africa. The genus Neuropteridiuvi is
important as an additional link between the Lower Gondwana floras
of India and South America on the one hand, and the South African
flora on the other. The genera Fsygmopkyllum and Bothrodendron
suggest interesting comparisons, from the point of view of geological
distribution, between South African and European Paleozoic floras.

* Feistmantel (89), pi. iv., fig. 2.

t These plants are described here for comparison with the plants of the same
age found within the limits of Cape Colony.

so Annals of the South African Museum.

FILICALES.

There can be little doubt that the genus Glossopterls should be
included in the Filicales. It is true we have not yet found fertile
specimens showing recognisable sporangia, but in addition to the
evidence furnished by the fern-like character of the sterile leaves,
the rhizome [Vertehrarla) " exhibits features characteristic of ferns.
As regards Gangaonopterls, which differs in no important respect
from Glossopterls, we may reasonably accept the view that it is also
a member of the FilicineaB.

Genus GLOSSOPTEEIS Brongniart.

Glossopteris browniana Brongn. var. indica.

Plate X., figs. 3, 4.

1828. Glossopterls Browniana Brongniart, Prodrome, p. 54.

G. Broioniana var. indica Brongniart, Hist. v6g. foss.,

p. 223, pi. Ixii., fig. 2.
1869. Glossopteris indica Schimper, Trait, pal. veg., vol. i.,

p. 645.
1896. G. indica Zeiller, Bull. Soc. G6ol. France, vol. xxiv. [3] ,

p. 366, figs. 11, 12, 13, pi. xvii., figs. 1-3.
1898. Glossopteris Browniana var. indica Seward, Quart. Journ.

Geol. Soc, vol. liv., p. 92.
1900. G. indica Potonie, Deutsch-Ost-afrika, vol. vii., p. 2, fig. 22.
1902. G. indica Zeiller, Obs. Sur. Quelques plant, foss. Lower

Gond. Mem. Geol. Surv. India, p. 8, pi. i.-iii.
1902. G. Browniana Arber, Quart. Journ. Geol. Soc, vol. Iviii.,

p. 5.
1902. G. indica, Zeiller, Flore fossile Tonkin (Atlas), pi. xvi.,

figs. 2-5.

The accurate identification or diagnosis of Glossopteris fronds, and
their separation into specific types, is, I believe, a hopeless task, and
I have purposely refrained from attempting a full synonomy of
Glossopteris browniana var. indica (or G. indica).

In a recent paper on fossil plants from New South Wales, Mr.
Arber t has given several references to literature in which records of
Brongniart's type occur. Important contributions have in recent

* Zeiller (96), p. 351. f Arber (02), p. 5.

Fossil Floras of Cape Colony. 81

years been made by Professor Zeiller, of Paris, to our kno\vleclji;e of
the venation characters of Glossoptcris fronds. In his account of
fossil plants from the neighbourhood of Johannesburg, Zeiller gives
drawings of the venation characters of the type-specimen ^= of CHos-
sopteris indica Schimp., and in his memoir on Indian plants f the
venation characters of this and other forms are admirably illustrated.
We still lack satisfactory information as to the taxonomic position
of the genus, although one is tempted to draw conclusions from
negative evidence.

The extraordinary alnindance of Glossopteris fronds monopolising
vast areas of sedimentary beds in different regions of India, South
Africa, Australia, and elsewhere, renders the absence of well-defined
and undoubted sporangia a striking fact. It may be tliat the fertile
fronds differed considerably in form from the sterile, or possibly
typical filicinean sporangia were not borne by the leaves of this
genus.

The most interesting consideration in regard to Glossopteris is its
geographical distribution. The announcement by Amalitzky I in
1897 of the occurrence of this plant in Permian beds of Vologda,
which has already been discussed by Zeiller in a paper published in
1898, § constituted an important contribution to our knowledge.

The Glossopteris leaves from Vereeniging are for the most part very
imperfectly preserved, and the venation characters are not sufficiently
distinct to describe in detail. In many cases the midrib is clearly
shown, but the secondary veins are eitlier invisible or exceedingly
obscure.

Plate X., fig. 1 (XXII.).

A broader form of leaf, 21 cm. long, agreeing in shape with speci-
mens of the type known as Glossoptcris browniana var. indica. The
midrib is distinctly shown, but hardly any trace of secondary veins
can be detected.

Glossopteris browniaxa Brongn. var. angustifolia.

Plate X., fig. 3.

I have previously adopted this designation for the narrow type of
Glossopteris fronds which Brongniart named G. angustifolia. The
specimens from Vereeniging are too imperfect to afford any guide
as to the advisability of regarding the narrower leaves as a distinct

* Zeiller (96), p. 307, tigs. 11, 12. t Ibi'l. (02).

: Amalitzky (97). § Zeiller (98).

82 Annals of the South African Museum.

species or including them under the comprehensive title Glossopteris
hroivniana.

Narrow Glossoj^teris leaves were found in the Vereeniging rocks in
association with Sigillaria, and described in 1897.*

Plate X., fig. 3 (XXII.).
This specimen, 20 cm. long, represents the long and narrower type
of frond comparable with Glossoj^teris angustifolia Brongn., but the
venation is too obscure to be represented in detail in the drawing;
the parallel lines in the figure are merely shading, but the darker
band in the middle of the frond indicates the presence of a distinct,
midrib.

Other specimens : Imperfect Glossoptcris fronds are seen in speci-
mens 1, 2, 3, XX., xxi., xxiv., xxv.

Genus GANGAMOPTEEIS, McCoy.

Gangamoptekis cyclopteroides. Feist.

Plate XIIL, fig. 5.

1869. Nceggerathia ohovata, Garruthers, Geol. Mag., vol. vi., p. 9,

pi. vi.
1876. Gangamopteris cijclopteroides, Feistmantel, Eec. Geol. Surv.

India, vol. ix., pt. iii., p. 73.
1879. Gangamopteris cyclopteroides, Feistmantel, Foss. Flor. Gond.

System, vol. iii., p. 12, pi. vii., ix.-xiii., xxvi., xxvii.
1899. Gangamopteris cyclopteroides, Feistmantel, Abh. k. bohm.

Ges. Wiss., p. 37, pi. iv., fig. 2.
1890, Gangamopteris cyclopteroides, Feistmantel, Mem. Geol.

Surv. New South Wales, Palaeont., No. 3, p. 131.

1894. Gangamopteris cyclopteroides, Kurtz, Eevist. Mus. de la

Plata, p. 13, pi. ii.

1895. Gangamopteris cyclopteroides, Zeiller, Bull. Soc. Geol.

France, vol. xxiii. [3], pi. x.

1896. Gangamopteris cyclopteroides, Bodenbender, Zeit. deutsch.

geol. Ges., vol. xlviii., p. 772.

1897. Gangamopteris cyclopteroides, Seward, Quart. Journ. Geol.

Soc, vol. liii., p. 323, pi. xxii., fig. 1, text-figure 1,
C. 324.

1898. Gangamopteris cyclopteroides, Seward, Quart. Journ. Geol.

Soc, vol. liv., p. 92.

• yewiud (97), p. 321, pi. xxi., fig. 4a.

Fossil Floras of Cape Colony. 8$

An inspection of the type-specimen of Carruthers' Brazilian leaf,'''
which he referred to Nmjgcrathia, convinces me that it should be
placed in Feistmantel's species Ganrjavioptcris cyclopteroidcs ; the
numerous and slightly spreading veins are undoubtedly connected
here and there by oblique anastomoses as in the typical Indian
leaves.

The association in Brazil of Neuropteridium validuiu (= Odontop-
teris plantiana Carr.) and Gangaviopteris]- (= Naggeratltia obovata
Carr.) is an interesting fact in view of the same association recorded
by Kurtz in the Argentine, and now known to occur in South Africa.
The specimens of Gangamoptcris included in the recently acquired
collection from Vereeniging are inferior in preservation to the leaf
that I previously figured from the same locality (British Museum
Collection, V. 3615).

Plate XIII., fig. 5 (A).

Part of a leaf showing numerous ol)lique and occasionally anasto-
mosing veins converging towards a central depression, where they
follow an approximately vertical course. The well-marked median
groove gives an exaggerated impression of a midrib, which in reality
is by no means as distinct as it is in a Glossoptcris frond.

FILICALES OR CYCADOFILICES?

The genus NeuropteridiiDU has been as a matter of course assigned
to the ferns, but we must recognise the fact that the Indian and
Southern Hemisphere fronds have not as yet been found in a fertile
condition. Another point of importance is the resemblance as
regards vegetative characters between this genus and NeuroptcriSy
a fact which suggests the possibility that Neurupteridi/iiii, like
Neuropteris and Alethopteris, may have been a plant in which
cycadean and filicinean characters were combined.

Genus NEUEOPTEEIDIUM, Schimper.

Schimper suggested the name Ncuropdcridluni as a subgenus of
Neuropteris in 1869 '[ for a type of frond from the Bunter of the

* Carruthers (69), pi. vi., tig. 1.

f See also Zeiller (95), pi. x.

+ Schimper (69), p. 447 : Schimper and Mougeot (44), pi. xxxvi.

■84 Annals of the South African Museum.

Vosges, which he named Neuropteridium gyandifoUa ; he defined
the subgenus as follows : —

" Prons simpliciter pinnata, pinnis integevrinis basi coarctatis,
infra mediam basim insertis ; nervo medio plus minusve distincto."

The long pinnate fronds of the type- species bore segments 4-5 cm.
in length and 2 cm. broad ; they differ in no essential features from
the leaf represented in pi. x., fig. 1.

In 1879 Feistmantel * published several drawings of large fronds
from Lower Gondwana rocks of India, which he at first designated
Neuropteris valida, but afterwards f Neuropteridium validuvi. These
Indian specimens, as Feistmantel pointed out, bear a close resem-
blance to the Triassic European species described by Schimper and
Mougeot.| The only evidence at present available, so far as I am
aware, in regard to the nature of the reproductive organs of fronds
included in NeuropAcridium is that afforded by specimens from the
Bunter rocks of the Vosges described by Schimper and Mougeot
under the name Crematopteris. Zeiller § considers that the fossils
so named are the fertile fronds of Neurop)teridium, and suggest a
comparison with certain recent species of Lomaria. There is little
doubt that Crematopteris of Schimper and Mougeot represents the
fertile condition of a Triassic species of Neuropteridium from the
Vosges sandstone — N. imhricatum S. and M. ; but the fertile pin-
nules exhibit nothing more than crowded dots or pits as shown in
the drawing (Schimper and Mougeot, pi. xxxv., fig. 3) without any
indication of spoi'angial structure. Granting the correctness of
Zeiller's view, and that the reproductive organs are of the nature
of sporangia, we are still without information as to the nature of the
fertile leaves of Neuropteridium grandifolium, the Triassic species
which so closely resembles N. validium. No fertile examples have
been described from the Talchir-Karbarbari beds, where the southern
species is particularly abundant, and it may be that — like the
northern Palaeozoic genera Alethopteris, Neuropteris, and other
fern-like leaves — Neuropteridium bore Gymnospermous rather than
filicinean reproductive organs.

In dealing with genera that are regarded as characteristic members
of the Glossopteris flora it is important to consider their possible
relationship or identity with northern Pala30zoic types. The genus
Neuropteridiwn is usually described as a fern confined to the Lower
Gondwana rocks of India and South America and to the Bunter beds
of the Vosges in Europe. We are now able to extend the range to

* Feistmantel (79, 79"), p. 10. \ Ibid. (81), p. 53,

t Schimper and Mougeot (44). § Zeiller (00), p. 109.

Fossil Floras of Cape Colony. 85-

Sontli Africa. There are, however, a few Nortliern Hemisphere
forms characteristic of Lower Carboniferous strata which exhibit a
very close agreement with Neurojjterulinni. The well-known Cardi-
optcris frondosa (Goepp.) from the Culm of Germany, as figured by
Stur,'" Schimper, I and others is characterised by large pinnate fronds
bearing semicircular entire segments in the lower part of the leaf
and longer and slightly lobed pinnules in the upper portion of the
frond, as in the African specimen represented in pi. x., lig. 1.

A portion of a pinnate frond almost identical with that showm on
pi. X. is figured by Fritsch from Thuringian Culm rocks as Archce-
opteris daivsoni Stur.]; A plant recently described by Grigoriew
from Upper Carboniferous rocks of Russia as Neuropteris cf. cordata
Brongn. var. densineura is also very similar to the African frond,
and should hardly be referred to a distinct genus, j It is probably
better, at least more convenient, to retain the two names Cardiop-
teris and Ncuropteridium ; but in some cases we must admit either
designation might be used with equal propriety as regards external
characters of the frond, irrespective of geographical distribution.
Cardiojyteris as at present used denotes fronds usually of Lower
Carboniferous age characterised by pinnules that tend to be more
orbicular and less deeply lobed than those of Ncuropteridium ; the
latter genus, moreover, with the exception of the Vosges species,
forms an element of the southern CUossoptcris flora, while the
former is northern in its range.

Neuroptekidium validum Feist.
Plate X., fig. 1.

1869. Odontoptcris Plav.tianaGiivvuthevs, Geol. Mag., vol. vi., p. 9,

pi. vi., figs. 2, 3.
1878. Neuropteris valida Feistmantel, Mem. Geol. Surv. India,

Foss. Flor. Gondwana System, vol. iii., p. 10, pi. ii.-vi.
1880. Neuropteridium validum Feistmantel, ibid. (1880), p. 81.
1894. Neurojiteridium validum Kurtz, Revista del Mus. de la

Plata, vol. vi., p. 11, pi. 1.
1896. Neuropteridiuvi validum Bodenbender, Zeit. Deutsch. Geol.

Ges., vol. xlviii., p. 772.

* Stur(7o), p. 41, pi. xiii., xiv.

t Schimper (69), p. 453, pi. xxxv. ; also Schlumberger and Schimper (62).
pi. XXV., p. .339.

I Fntsch (97), pi. i., tig. 3, p. 90. § Grigoriew (98), pi. iv.

S6 Anncds of the South African Museum.

Frond long and linear, pinnate, rachis strong, bearing pinnules
which in the lower part of the frond are entire and more or less
semicircular in form, and gradually pass into longer and lobed seg-
ments as we ascend the rachis. The longer pinnules, which may
reach a length of 6 cm. or more, are not attached by the whole of
the base, like the stouter and broader segments in the basal portion
of the frond, but the basal lobe of the upper edge of the segment is
free from the rachis. Apex of pinnules bluntly rounded. Veins
spreading, curving towards the edge of the pinnules with repeated
dichotomous branching, and converging in the longer segments to
form a fairly distinct midrib in the lower part of the lamina.

The discovery of Neuropteridium at Vereeniging by Mr. Leslie, to
which allusion has been made in a recently published article,*
affords a very interesting confirmation of the striking similarity
between the vegetation of the Talchir-Karharberi series of India and
that of South America and South Africa. There can be no reason-
able doubt as to the identity of the Indian species and the plant
from Vereeniging. Similarly Carruthers' Odoiitojyteris idantiana
from Eio Grande do Sul, Brazil, is undoubtedly the same type, and
Kurtz has already recognised the identity of Argentine specimens
with Peistmantel's species. The type-specimens of Carruthers'
plant from Brazil are in the British Museum Collection, and an
examination of them confirms my opinion as to the identity of the
Bi-azilian, African, and Indian fronds.

Plate X., figs. 1, lo, Ih (XXII.).

Fig. 1 is drawn slightly less than natural size. Frond 54 cm. long.
The basal portion is probably close to the actual termination, but the
apex has not been preserved. The lowest pinnule, which is almost
orbicular in outline, is approximately 2-5 cm. in breadth and 2 cm.
in length ; the spreading and forked veins are faintly indicated. The
fourth segment from the bottom of the rachis is 2-6 cm. long and
2"9 cm. broad, the margin appears to be entire, but the upper edge
is flatter than the more strongly arched lower margin. The next
pinnule is slightly larger, and shows the beginning of a subdivision
of the lamina into broad lobes. Pinnule number 7 reaches a
length of 5 cm., and has well-marked lobes with the Ncuroptcris type
of venation. Tliis segment is shown very slightly larger than

* Seward (02).

Fossil Floras of Cape Colony. 87

natural size in fig. la. The ninth pinnule is 6 cm. lonj^ ; pinnule
number 13 is shown approximately natural size in fig. 16. Towards
the apical margin of the frond the segments gradually decrease in
size, precisely as in the Indian examples described by Feistmantel.

Plate X., fig. 2 (XXVI.).

A piece of the rachis near the base of a frond showing
semicircular and orbicular segments. The veins are not seen.
Associated with Glossoptcris fronds.

LYCOPODIALES.

Genus (?) BOTHEODENDEON Lindley and Hutton.

BOTHEODEXDRON LESLII sp. UOV.

Plate XL, figs. 1, la, Ih, 4, 5, 6.

One of the most striking differences between the Upper Carbo-
niferous and Lower Permian vegetation of the Northern Hemisphere
on the one hand, and the Indian and Southern Hemisphere Glossop-
tcris flora on the other, is the absence in the latter of many of the
most abundant and characteristic northern genera such as Galaviitcs,
Lepidodendron, Sigillaria, and others. In recent years, however,
instances have been recorded of the admixture of northern Palaeozoic
types with members of the Glossopteris flora : this is notably the
case in South America and South Africa. Prior to the appearance
of the Glossoptcris flora in Australia there existed in that continent,'''
as in South America, t certain species of plants closely resembling
European types — a fact pointmg to a greater uniformity in the vege-
tation of extreme northern and southern latitudes during the Lower
Carboniferous period than prevailed at a later stage. Lower Carbo-
niferous plants have been recorded by Carruthers,^ Feistmantel,?
Etheridge,j| and others from Australia, which exhibit a more or less
close agreement with certain species in the Northern Hemisphere.
Such genera asBhacoptcris, Bothrodendron, d^ndi Lepidodendron afford
instances of a close correspondence between the extreme northern
flora of the Arctic regions and that which existed in the far south.

* Feistmantel (1)0). t Kurtz (94). See also Zeiller (97).

+ CaiTuthers (72). i^ Feistmantel (90).

il Etheridge (90) (90^^) ; Jack and Etheridge (92).

88 Annals of the SoutJi African Museum.

From South America Lower Carboniferous plants are recorded by
Szajnocha ''•'■ and Kurtz,! which, hke those from Austraha, point
to the ahiiost world-wide occurrence of a vegetation of uniform
facies.

It has often been stated that such European types as Calamites,
Lrpidodcndron, and others, also occur in South Africa, but apart from
the recently recorded instaiices of the association of Sigillaria with
CTlossopteris, &c., it appears to be the case that the statement qu.oted
))y one author from another is based in part on incorrect determina-
tion, and largely on a confusion of localities, English or American
Coal-Measure fossils having been erroneously assigned to a South
African locality. In 1871 Grey ;[ referred to the occurrence of
Lcpidodcndron and Sigillaria in the Stormberg district, but Professor
Eupert Jones § expressed his belief that the plants in question were
either American or English. I have no doubt, as Eupert Jones
pointed out, that Grey's specimens were not originally obtained from
Africa ; this belief is founded on an examination of several specimens
in the British Museum (V. 235, V. 2390, V. 3267), e.g., Calamites,
Calamocladus, Lepidodendron, and others, said to have been obtained
from Stormberg beds.

We may next consider what authentic evidence is available bear-
ing on the occurrence of Lepidodendroid plants in South African
rocks. A few years ago Mr. David Draper, of Johannesburg, for-
warded a collection of plants to me from several localities in the
Transvaal. Among them were several specimens from Vereeniging,
which I described [[ as Sigillaria hrardi Brongn., a species charac-
teristic of a high horizon in the North Hemisphere Coal-Measures,
and met with also in Permian rocks. This Sigillaria was found in
association with Glossoptcris, Ciangamoptcris, and Na'ggerathiopsis,
in strata assigned to a Permo-Carboniferous age. Before dealing
with the recently discovered plant represented in pi. xi., reference
may be made to a few specimens in the British Museum which
afford additional illustrations of South African Lycopodiaceous
plants. The fossil represented in text-figure 8 (V. 236) was obtained
from a locality spoken of as Atherstone Quarry, Kowie. It consists
of a piece of flattened stem, 8 cm. long and 2*5 cm. broad, bearing
spirally disposed circular protuberances and curved linear append-
ages, presumably leaves, of wdiich some are shown in attachment to
the edge of the impression. The preservation is not sufficiently

* Sza,)nocha-{91). t Kurtz (94). + Grey (71).

§ Jones (84). |1 Seward (97).

Fossil Floras of Cape Colony.

89*

good to allow of specific determination, l)ut there can l)e little doul)t
that the fragment belonged to a Lepidodendroid plant.

Text-figure 8.

Lepidodendroid stem. (Nat. size.)
From a specimen in the British Museum (V. 236).

A few other fragments in the Museum Collection (V. 2424) are
stated to have been obtained from a locality near Port Alfred, Kowie
mouth.

An obscure specimen (V. 242.3), which may represent a fragment
of a Lepidodendroid plant (? Botltrodendron), was found in a boulder
from the Prince Albert Eange, possibly from the Wittebei'g. The
Museum Collection also includes similar fragments from Lower
Albany (V. 240), which may be of Devonian age.''- In addition to
these fragments there are two imperfectly preserved sandstone casts
obtained by Mr. Draper from Zwartkoppies, Vredefort, in the Orange
River Colony, which are no doubt partially decorticated stems of
some Lepidodendroid plant.

It is possible that the specimen represented in text-figure 8 and
that from the Prince Albert Eange may be of Devonian age. In
this connection it may be mentioned that a small specimen from
Witteberg included in the collection recently sent to me (No. 53a) is
in all probability a fragment of a Lepidodendroid stem.

* An example of a Lepidodendroid plant is referred to hy Mr. Draper in his paper
of 1897, p. ;3ia.

•90 Annals of the South African Miiscmn.

The specimen represented in pi. xi., fig. 1, does not show the
•surface features of the plant sufficiently clearly to enable one to
determine its affinity with certainty ; it is, I believe, a Lepidoden-
droid species, but we have no satisfactory indication of the nature of
the appendages borne on the spirally disposed scars. In all prob-
ability, however, the appendages were leaves, and the fossils present
the appearance of aerial rather than subterranean organs.

My friend Professor Zeiller has drawn my attention to some frag-
ments figured by Feistmantel * in his memoir on the Damuda and
Panchet beds of India, referred to as pieces of the stem of a coniferous
plant, probably Bhipidoims, which bear a resemblance to the
Vereeniging fossils. We may also suggest a comparison with a
specimen figured by Schmalhausen from Petschora as a piece of the
stem of Bhipidopsis ginkgoidcs.]' Feistmantel's specimens are too
indistinct to determine, and there is no satisfactoiy evidence in
support of his determination. The Petschora fossil seems to agree
more closely with the Vereeniging stems, and may be Lepidoden-
droid. I am indebted to Professor Zeiller for suggesting that the
fossils represented on pi. xi., figs. 1, 4, 5, 6, may be gymnospermous
stems, their scars being very similar to the leaf-scars of the branches
of recent species of Abies. It must be admitted that the surface
features of the African specimens do not conform in detail to those
of Bothrodcndron, but in all probability we have to deal with par-
tially decorticated stems, which would not exhibit the parichnos
scars nor the characteristic sculpturing of the bark. The apparently
dichotomous branching shown in fig. 4 is, I think, an additional
argument in favour of assigning the specimens to the Lycopodiales.

In his recent and very important memoir on Upper Devonian
plants of Bear Island, Professor Nathorst figures nume5'ous examples
of a Lepidodendroid plant which he identifies as Bothrodcndron
kiltorkensc (Haugh.).:[: Haughton's species, Guclostigma kiltor-
kense,l from the Kiltorkan beds of Ireland, was placed in the genus
Bothrodcndron by Kidston || as the result of an examination of a
series of specimens in the Dublin Museum. Some of these Arctic
specimens bear so close a resemblance to the South African plant
that one might be almost tempted to regard the two as identical
species. My friend Professor Nathorst, to whom I forwarded a
drawing of the African stems agrees with me that they may belong

* Feistmantel (81=), pi. xlvii.rt, figs. 5-7.

t Schmalhausen (7i)), j)]. viii., fig. 12.

+ Nathorst (O'i), p. ;n, pi. x.-xiv.

§ Haughton (GO). |i Kidston (8<l), p. (5(5.

Fossil Floras of Cape Colony. 91

to a species of Bothrodcndron, but considers them specifically distinct
from B. kiltorkcnsc. The small forked specimen shown in fig. 4 may
be matched almost exactly with Nathorst's figs. 1, 2, 5, 6, 10, and 11
in pi. xi. of the Bear Island Memoir. The larger and more perfect
specimen from Yereeniging (pi. xi., fig. 1) also agrees closely with some
of Nathorst's examples. In spite of the close superficial agreement, I
believe the African plant to be a distinct species — a view which my
friend Mr. Kidston also unhesitatingly adopts. In the stems from
South x\frica I am unable to detect any trace of the two lateral scars,
representing the parichnos, sliown in some of Nathorst's figures of
the leaf-scars of Bothrodcndron kiltorkense ; the absence of the
parichnos scars may be due to the partial decortication of the stems.
Another plant which may be compared with the African species is
Cyclostigma australe, as figured by Feistmantel ■■'• from Devonian
rocks of New South Wales. This Australian type is possibly
identical with Haughton's species. The variation in the arrange-
ment of the scars is in accordance with a reference of the fossils to
the Lycopodiales ; striking examples of Sigillaria with leaf-scars
closely approximated in certain regions of the stem, and farther
apart in other regions, have been figured by Kidston, Zeiller, and
other authors.

The Yereeniging plant appears to agree as regards the size, form and
arrangement of the scars with the genus Bothrodendron (including
under this term Cijclostujma) more closely than wdth other members
of the Lycopodiales. I purpose to designate the specimens Bothro-
dendron leslii.

Plate XL, figs. 1, la, Ih (XXL).

A piece of a stem, 29 cm. long, and varying in breadth from 1-5 to
2 cm. The small piece at the upper end — as shown in the drawing
— is at a higher level than the rest, and shows the almost circular
scars as prominences with a central depression. Below this the
scars occur as depressions wdth a central umbo, and in some of them
one is able to detect a small pit in the centre of the umbo (la), the
pit no doubt marking the position of the vascular bundle, which
supplied the spirally disposed appendages. In all probability the
uppermost piece of the specimen is the impression of the outer, but
partially decorticated, surface of the organ, while the remainder of
the specimen represents the inner face of the compressed cortex, the
internal tissues — wood and pith — having been removed. At c the

* Feistmantel (1)0), pi. xi.

92 Annals of the South African Museum.

scars become more crowded and the surface shows signs of wrink-
ling— no doubt the result of contraction on drying. Lower down the
scars again become more scattered, and later, h, the more crowded
disposition recurs. The enlarged drawing of the region h (Ih) shows
more clearly the nature of the scars and their arrangement : in many
of them the preservation is not sufficiently good to show^ the vascular
bundle pit in the central umbo. Below^ the region h the surface
features become more obscure.

Plate XI., fig. 4 (B).

This fragment is of interest as showing the dichotomous branch-
ing ; the scars are precisely like those of the larger specimen, but.
their preservation is less perfect.

Plate XL, fig. 5 (XX.).

An obscurely preserved and wrinkled specimen, similar to xxi.

(fig- !)•

Plate XL, fig. 6 (XX.).

A larger branched fragment, bearing scars like those of xxi..
(fig. 1), but showing the surface features much less clearly. There
are indications here and there of impressions suggesting narrow
linear, leaf -like appendages lying on the stem, but the preservation
is too poor to enable one to recognise any certain traces of
appendages.

? GINKGO ALES.

Gexus PSYGMOPHYLLUM Schimper.

This genus was instituted by Schimper in 1870 " for leaves of
Permian and Upper Carboniferous age previovisly referred to Ncegge-
rathia. He defined it as follows : —

" Folia pinnatisecta, pinnis erecto patentibus, e basi valde angus-
tata flabelliformibus, longitudinaliter flabellatim plicatis, plus minus
profunde pinnatisectis, vel margine lobatis sen crenatis ; nervis
plusies dichotomis, erecto-radiantibus. Vernatio foliorum verti-
caliter involuta."

Psyrjmophyllum is placed by Schimper among the Cycads, while
later authors i have prefexTed to regard the lobed and wedge-shaped

* Schimper (70), p. 192. \ yapoita (84), p. 230 ; Zeiller (00), p. 251.

Fossil Floras of Cape Colony. 93

leaves as probably those of a member of the Ginkgoales. We have,
however, no evidence beyond such as is afforded by the not very
close resemblance of the leaves to those of Ginkgo that can be
accepted as in any way deciding the position of Schimper's genus."-

The genus Ginkgophijllum was adopted by Eenault and Saporta f
for leaves which cannot be distinguished from Psj/gmopiij/llmn, but I
see no reason for substituting the more committal term for Schimper's
older designation. Feistmantel's Eiirypliylluvi, wJiitticnmnt, I from
the Karharbari Coal-field may be compared with Psygmophyllam as
regards the tapered base of the leaves and tlieir arrangement on
the stem, but the Ginkgo-Yike form of the iDilobed lamina affords a
distinctive feature of the latter genus.

PSYGMOPHYLLUM KIDSTONI, Sp. nOV.

Plate XII., fig. 1.

Vegetative shoots woody, bearing spirally disposed leaves. Leaves
wedge-shaped, reaching a length of 13 cm., usually divided by a deep
median sinus into two narrow wedge-shaped lobes, truncate distally
and tapering gradually to the proximal end of the lamina. Leaves
sessile, attached to the axis by a narrow base. The lamina is
traversed by numerous spreading and occasionally forked veins
following a course parallel to the edge of the leaf.

Organs of reproduction unknown.

Leaves of the Psygmophyllum type have hitherto been recorded
from the following regions and geological horizons : —

Psygmophyllum Jiabcllatum (L. and H.),; Coal-Measures of New-
castle-upon-Tyne, England.

Psygmophyllum icilliamsoni, Nath,|| Culm beds of Spitzbergen.

Ginkgophyllum flahdlatum (L. and H.), Eenault, Cours. Bot. Foss.
1881, p. 65, pi. vii., fig. 5.

Psygmophyllum expansum (Brongn.),'i Permian of Russia.

Psygmop)liyllum grasscti (Sap.),^'"'' Permian of Lodeve.

Ginkgophyllum minus, Sandb.i i Coal-Measures or Permain of the
Black Forest.

* Seward and Gowan (00), p. 136.

t Eenault (81), p. 65; Saporta (84), p. 2.30.

I Feistmantel (79), pi. xxi. § Lindley and Hutton (32), pi. xxviii., xxix.

II Nathorst (94), pi. ii., figs. 1, 2.

U Murchison (45), vol. ii., p. 9, pi. E. See also d'Eichwald (55), pi. xiii., fig. 17,
.and Sehmalhausen (87), p. 18, pi. iii., iv.

** Saporta (84), p. 228, pi. clii., fig. 2. ft Sandberger (90), p. 101.

94 Annals of the South African Museum.

Two specimens of Psygmoplujllum fiahcllatum in the British
Museum from the Coal-Measures of Newcastle (England) bear so close
a resemblance to the South African plant that I am unable to detect
any well-marked differences. One of these specimens (40578) is a
large leaf tapering gi^adually to a narrow basal region with a lamina
15 cm. long and 11-5 cm. broad, agreeing in size and shape with the
example shown in pi. xiii. The other specimen (38927) shows an
axis like that represented in pi. xiii. bearing portions of two long
wedge-shaj)ed leaves. The collection includes a similar specimen,,
but less clearly preserved, from the Province of Perm, in Eussia,
referred to Psygmopliyllum cxpansnm Brongn.,''' a type which is
probably closely related to the species of Lindley and Hutton. f
Nathorst's :|: Psygniophylluni williamsoni from Spitzbergen bears a
resemblance to P. hidstoni, and is of interest as coming from
a lower horizon than the other examples of the genus. Several
fossils have been figured by Schmalhausen from the Permian of
Eussia under the generic name Psyginophyllum, and some of these
undoubtedly represent species bearing a close resemblance both tO'
P. flabeUatnm (L. and H.) of the English Coal-Measures and to the
Vereeniging plant ; this is the case with some at least of the speci-
mens referred by Schmalhausen to Brongniart's species Psygmo-
phijllum expansum.l On the other hand, such forms as that
represented in his fig. 11, pi. iii. — referred to P. cuneifolium (Qvongn.)
— can hardly be accepted as species of this genus. The specimen
named by Schmalhausen Baicra gigas |] is in all probability a leaf of
Psygmopliyllum, and may indeed be identical with the African plant.
The leaves from Eussia described by Schmalhausen as Piliipidopsis
glnkoides *'■■ differ from those of tlie African plant in being compound,
luit the individual leaflets are very similar to the leaves represented
in pi. xiii. Under the name Arcluwptcris archctypus the same author
has figured leaves from Upper Devonian age in Eussia*" which may
be compared with Psygmopdiyllum Jcidstoni ; the habit of the branch
as shown in one of Schmalhausen's figures (pi. ii.) appears to be
identical with that of the African plant. Nceggcrathia foliosa Sternb.,
a species chai'acteristic of the Eadnitz Coal-Measures, Bohemia, and
described by Feistmantel from Silesia,! f also bears a resemblance to

* Brononiiut in Murchison (45), pi. E.

f Lindley and Hutton (.32), pi. xxviii.-xxix.

\ Nathorst (94), pi. ii., figs. 1, 2.

§ Schmalhausen (87), pi. iii., fig. 10; pi. iv., figs. 1-G.

II Oj). cit., pi. v., fig. 10. II Schmalhausen (79), pi. vi.

** Ibid. (94), p. 23, pi. ii. ft Feistmantel (75), pi. v., fig. 1.

Fossil Floras of Cape Colony. 95-

the Vereeniging plant as regards the form, venation, and manner of
attachment of the leaves. As an example of a plant from a higher
geological horizon not unlike the African type, reference may he
made to Noeggerathia vogesiaca Bronn from the Eaibl Ehsetic beds,
but Bronn's plant has a broader axis, and the longer leaves are
disposed in two rows like the pinnae of a Cycadean frond.'''

Plate XII., fig. 1 (XXIII.).

The woody and fairly stout axis reaches a length of 34 cm., the
leaves appear to be arranged spirally, and in some of them the
attachment to the axis is clearly shown. The venation is faintly
indicated, but sufficiently obvious to show that the lamina is
traversed by numerous veins following a course approximately
parallel to the edges of the leaf. The leaf a, which is incomplete,
and may indeed represent half the original lamina, shows the gradual
tapering towards the base. Leaf b is still attached to the axis, and
illustrates the bilobed character, the wedge-shaped lamina being
divided by a deep median sinus into two halves, one of which has
become pressed down to a slightly lower level in the matrix of the
rock than the other. In the incomplete and lolied lamina of leaf c
the veins are fairly distinct. A portion of leaf d is seen in direct
continuity with the axis. The lamina of leaf e is almost perfect, and
shows the bilobed character very cleai'ly.

CORDAITALES.

Genus NCEGGERATHIOPSIS, Feistmantel.

Feistmantel's generic name instituted in 1879 j is another illustra-
tion of the influence of geographical distribution on nomenclature.
Had the leaves referred to this genus been found in European
Palaeozoic rocks there can be little doubt that they would have been
described under the name Cordaites. Professor Zeiller | has drawn
attention to slight differences as regards venation between the leaves
of the two genera, and in fragments of cuticle detached from a South
African specimen the disposition of the stomata appears to differ to
some extent from that in Cordaites. Kurtz § in a recent note has
incidentally referred to the more spreading arrangement of the veins

* Bronn (o8), pi. vi. t Feistmantel (79), p. 23.

+ Zeiller (96), p. 23. See also Feistmantel (90). § Kurtz (03), p. 25.

■96 Annals of the South African Musenin.

as a distinguishing character of Nceggerathiopsis, but the concensus
-of opinion favours the view that these northern and southern leaves
with parallel or slightly spreading veins belong to closely allied
genera, and should both be included in the Cordaitales. In his
recent memoir on Lower Gondwan a plants Zeiller " expresses himself
■strongly in favour of regarding Noeggerathiopsis hislopi (Bunb.) as a
species of the Cordaiteae, and very closely allied to Cordaites. Solms-
Laubach,! Krasser,]: Arber,§ and other authors adopt a similar
view.

The leaves of Heer's genus Phoenicopsis, which I believe to be
identical with Zeugophyllitcs, differ from the broader leaves of
Noeggerathiopsis in the al)sence of the frequent dichotomous branch-
ing of the veins. As Zeiller says, we cannot resist the inference
that the Gymnospermous seeds found in association with Nceggera-
tJiiopisis in the Karharbari beds may, like the European Cardiocarpits
seeds in association with Cordaites, \v^yq been borne by this southei'n
member of the Cordaiteae. 1| The leaf from Tonking named by
Zeiller 'I Nceggerathiopsis hislopi bears so close a resemblance to the
Stormberg plant Phanicopsis elongata that I am inclined to think
the latter generic name more appropriate than Noeggerathiopsis.

Nceggerathiopsis hislopi (Bunb.).
Plate X., fig. 5 ; Plate. XIIL, figs. 2-4.

1861. Nceggerathia Hislopi, Bunbury, Quart. Journ. Geol. Soc,

vol. xvii., p. 334, pi. x., fig. 5.
1879. Noeggerathiopsis Hislopi, Feistmantel, Foss. Flor. Gondwana

System, vol. iii., p. 23, pis. xix., xx., xxviii., xxix., xxx.
1879. ? Rhiptozamites Goepperti, Schmalhausen, M6m. Acad, imp.,

St. Petersburg, vol. xxvii. [7] , p. 32, pi. iv., figs. 2-4 ;

pi. vii., figs. 23-27.
1887. lihiptozamitcs Goepperti, Schmalhausen, Mem. Com. Geol.,

vol. ii., No. 4, pi. v., fig. 6.
1890. Cf. Nceggcratliiopsis spathnlata and N. media, Feistmantel,

Mem. Geol. Surv., N.S. Wales, p. 154, pi. xxi.,

figs. 3-5.
1897. Noeggerathiopsis Hislopi, Seward, Quart. Journ. Geol. Soc,

vol. liii., p. 322, pi. xxi., fig. 4&.

* Zeiller (02), p. ai. f Solms-Laubach ('Jl), p. 110.

\ Krasser (00). ij Arber (02).

II Zeiller (02), p. a2. II Ibid. (82), pi. xii., fig. 13 (02=), pi. xl.

Fossil Floras of Cape Colony. 97

1894. NoRcjgerathiopsis Hislopi, Kurtz, Eevist. Mus. de la Plata,

p. 15, pi. iii., figs. 3, 4 ; pi. iv., figs. 1-3.
1896. Nosggerathiopsis Hislopi, Zeiller, Bull. soc. g^ol. France,

vol. xxiv. [3], p. 372, figs. IG, 17, and pi. xviii.,

figs. 6-9.
1898. Naggerathiopsis Hislopi, Seward, Quart. Journ. Gaol. Soc,

vol. liv., p. 93.
1900. ? Nosggerathiopsis Hislopi, Potonie, Deutsch-Ostafrika,

Bd. vii., p. 7.
1902. 1 Nceggerathiopsis Goepperfi, Arber, Quart. Journ. Geol. Soc,

vol. Iviii., p. 17, pi. i., figs. 1, 2.

Leaves broadly linear, varying considerably in size and shape,
traversed by numerous parallel veins exhibiting dichotomous branch-
ing, more especially in the basal portion of the lamina, but occasion-
ally also in other parts of the leaf. In the longer leaves the veins
follow a straight course through the greater part of the lamina, but
in the shorter and obovate leaves the veins tend to diverge towards
the edges in the upper part of the lamina. The epidermal cells have
rectilinear walls.

The Eussian fossils originally referred by Schmalhausen =•' to
Bhiptozamitcs from strata which are said to be of Jurassic age, but
which have been more recently demonstrated by Zeiller | to be
Palaeozoic, are no doubt very closely allied to Nceggerathiopsis
Jiislopi. Bhiptozamites gocppcrti is recorded also from Eussian
Permian rocks by Schmalhausen, and the leaf figured in his pi. v.,
fig. 6 X can hardly be specifically distinguished from the smaller
forms of Noeggeratliiopsis hislopi from Vereeniging. Peistmantel §
and Zeiller || have both drawn attention to the resemblance between
Bhiptozamites and Cordaites, and there can be little doubt that
Schmalhausen's plant should be placed in the Cordaiteae and regarded
as generally identical wdth Naggerathiopsis or Cordaites. Arber*
has described imperfect leaves from Australia which, as he observes,
are probably identical with the Eussian type ; he notes the close
resemblance between Bunbury's species N. hislopi and N. gocppcrti,
but prefers to retain both specific names.

Plate XIIL, fig. 4 (XXIV.).
Part of the torn lamina of a leaf, incomplete at both ends. The
leaf is traversed by slightly spreading and crowded veins which

* Schmalhausen (79). t Zeiller (96=^). I Schmalhausen (87).

§ Feistmantel (90), p. 153. 1| Zeiller (96-^), p. 475. *: Arber (02), p. 17, pi. i.

7

98 Annals of the Soutli African Museum.

occasionally bifurcate. The lamina is curved so as to present a
convex surface ; it extends distally into the sandstone, so that the
apparently triangular apex does not represent the actual termination
but merely the boundary of the exposed part of the specimen. This
leaf may be compared also with Euryi^hyllum, as figured by Feist-
mantel from India.

Plate XIII., fig. 2 (XXI.).

Part of a leaf 12'5 cm. long and 4 cm. broad. The venation is not
very clearly shown, but the specimen is of interest on account of its
striking resemblance to a leaf of Cordaites.

Other si^ecimens : Fragments occur on specimens xx., xxi., xxii.,
xxiv. One clearly preserved impression (xxi. 6) represents a broad
linear leaf like that shown in pi. xiii., fig. 2, but with a naiTow basal
termination.

Plate XIII., fig. 3 (XXV.) (1 nat. size).

A torn leaf 10 cm. long, showing the original form of the lamina.
The slightly spreading veins appear to be rather farther apart than
those in the specimen shown in fig. 4, but this need not necessarily
indicate a specific difference, as some at least of the finer ridges
in fig. 4 may represent hypodermal strands of mechanical tissue such
as are known to occur in the leaves of Cordaites.

Plate X., fig. 5 (XX.).

Part of a leaf showing the veins more clearly. The narrow end of
the fragment does not represent the actual apex, as the lamina
extends further into the rock. The veins are numerous and more
oblique to the margin than in the other figured specimens.

PLANT OF DOUBTFUL POSITION.
A. — Ehizobie or Eggt. '

Plate XII., fig. 2 (XX.).

This imperfect specimen, represented natural size in the drawing,
may be a rhizome giving off numerous branched roots, but it is
impossible to speak with confidence as to its precise nature. It is
perhaps not improbable that the fossil may be part of the rhizome,

Fossil Floras of Cape Colony. 99

Avith roots, of Glossopteris ; a comparison may be suggested with the
figure of Vcrtebraria {Glossopteris) published by Bunbury."

B. — Stem.
On specimen xxi. occurs a flattened impi-ession of a woody stem
55 cm. long and 4 cm. broad, showing a well-defined pear-shaped
branch-scar. In the absence of petrified woody tissue I am unable
to offer an 3' opinion w^orthy of expression as to the affinity of the
fossil.

III.— CONCLUSION.

The plants from Worcester are unfortunately too few in number
and too fragmentary to enable one to make a definite statement as to
the age of the beds, but so far as the evidence goes it points to a close
resemblance of the flora with that which is more abundantly repre-
sented in the plants of the Vereeniging sandstones. I shall, therefore,
confine myself to the question of the age and correlation of the
Vereeniging species. It has been pointed out in the section of this
paper dealing with Bothrodendron that the supposed occurrence of
European species of Lcpidodcndron, Calamitcs, and Pccopteris in
South African rocks is in all probability erroneous.

The European Permian and Coal-Measure species Sigillaria hrardi
Brongn. is represented by satisfactory specimens obtained from the
Vereeniging rocks, and we are now able to add a second example of
the Lycopodiales to the list of South African plants, but there is
absolutely no proof that the Northern Hemisphere Permo-Carboni-
ferous flora existed in Africa south of the Zambesi in Portuguese
East Africa.! Without attempting a complete survey of the literature
dealing with South African plants, which there is good reason for
assigning to a Pal£eozoic horizon, we may briefly consider the general
facies of the vegetation of which remnants have been recorded.
Leaving out of account such fragments as cannot be satisfactorily
determined, w^e may regard the following types as a representative
series of the older vegetation of the Transvaal, Cape Colony, and other
districts. Numerous specimens of Glossopteris have recently been
obtained from Zululand and in Natal this genus and other Palaeozoic
plants are known to occur. Potonie has also recently described a few
forms from Portuguese and German East Africa, but the greater
number of the plants mentioned in the following list are from the
neighbourhood of Johannesburg, Vereeniging, and other localities in
the Transvaal.

* Bunbmy (Gl), pi. xi. t Zeiller (S3).

100

Annals of the South African Museum.

P? CO

5 H
^^

Pi H
». to

5^

§^

P o

wo

:3

« to

H P<

FUCO

O

Pi

P

« P5

< M

-2 g^
^.^^

^ C c3 S

fa, . r- aj
2 i£,_g '^
£-. s g o

§111

"o

(S . S ^-"S.

33

O

es sp. e
'ozamites (?
(')-f i as figui
z Joseph L
on and Te

&

M .

0) c

2.2

•S Sh ^ S p.

^

s ^

£'M^^^ %

W ;=

1^%^'^

&

.o

o

o

tm

s -

OQ e

'§S

,

>-,

.o

05

al

-Ti

0)

1^

C3

SoS

t3

c«

Si

m e S

b^

02

:^ '^ CO

rw -^ ^ -w O) .

^ Si "Su -r o ^ cj

^ S M .^ CD ^ '-^

O ~ ,_ .^ si; O <D

'S S c 2 oj o rj

^- C3^ tH o<!

e _C2 ra

s.g:^ s,

I^dSJ

; o c2 ce

O

~ c: ta •" '^

CO rC 2 ^ '5 ^

a^ =:h§^K^ 2-«^fl

« S g

t/2 §-^

«

■3

2i

00 eii CO

Fossil Floras of Cape Colony. 101

Tlie conclusion to bo drawn from the Vereeniging plants is that
they belong to a flora which flourished in South Africa, India,
South America, and Australia during some portion of the Permo-
Carboniferous epoch. On the whole, it would seem probable that
the age of the plant-beds corresponds most nearly with the Upper
Carboniferous period as represented in Europe. It is of necessity
■difficult to attempt to express the geological age or homotaxy of
South African beds in terms of the geological chronology of the
Northern Hemisphere, but the close correspondence of some of the
Vereeniging types with Indian and South American species points to
their correlation with the Karharbaii beds of the Lower Gondwana
system. The occurrence of such types as Siglllaria, Bothrodcndron,
and Psygmophyllum shows a closer correspondence between the
South African flora and that of the Northern Hemisphere than occurs
in the Indian vegetation ; we have evidence of an overlapping or
commingling of the northern and southern botanical provinces " in
South Africa and in South America that is not afforded by the Lower
■Gondwana floras of India and Australia.

* Zeiller (97).

102 Annals of the South African Museum.

lY.—WITTEBEBG SEBIES.

I DO not propose to enter into a discussion as to the stratigraphy
of the Witteberg beds : * the numerous examples of Sjnrophyton which
they have afforded do not, in my opinion, supply a clue to their
geological age. A small and obscure fragment of a Lepidodendroid
plant, which is the only recognisable and undoubted fossil plant
included in the collection, would be in accordance with the reference
of the series either to the Carboniferous or Devonian period. Con-
sidering the relation to the Witteberg series to the rocks containing
the species described in Part III. of this Memoir, it would seem
probable that they belong rather to the Devonian than to the
Carboniferous epoch.

LEPIDODENDEOID PLANT.

(53a) A small and obscure impression on a quartzose rock showing
crowded and spirally disposed projections. Too small and imperfect
to determine. From Eooiberg, Breede Eiver.

EOOTS (?)

The quartzose rocks in which the Spirophyton occurs contain also
some radially disposed and comparatively deep grooves, which differ
from the ridges and grooves of Spiropliyton in being straight and few
in number ; they probably represent roots which spread into the
sand from a central axis. The preservation is not sufficiently good
to enable one to hazard a suggestion as to the nature of the plant to-
which they belonged, but they certainly dilfer from the typical Spiro-
'pliyton, and should not be included under that head.

(44a) A large block of quartzite with a funnel-shaped depression on
one face, the sides of which are marked by irregular deep grooves
which may have been formed by radially disposed roots.

(45a, 47a, 50a) A smaller specimen showing at one end half of a
conical elevation bearing radially arranged grooves ; on the face of
rock exposed in a longitudinal section through the apex of the conical
elevation one sees indications of a central line or axis, from which
iron-stained bands extend through the matrix. This specimen,

* 8ee Feistniantel (81)) ; Annual Eeport (90).

Fossil Floras of Cape Colony, 103

like 44a, 45a, and 47a, does not exhibit the typical Spirophyton
characters.

Genus SPIROPHYTON, Hall.
(Probably of mechanical and not vegetable origin.)

Among the specimens from Cape Colony there are a few large
blocks of hard quartzose rock containing well-marked impressions of
the problematical fossil Spirophyton from the Witteberg sandstone,
1 mile N.E. of Touws River Station, from Zout Kloof, 3 miles E.S.E.
of Ladismith and from the north bank of the Touws River at Letta's
Kraal.

The Reports of the South African Geological Commission contain
several references to the abundance of Spirophyton cauda-galli in
the Witteberg quartzites ; ■■'- it is mentioned as being particularly
abundant towards the top of this series, and is regarded as a
convenient and apparently trustworthy means of recognising the
Witteberg beds.

In 1842 Vanuxemi pul^lished a figure of a " fucoid," which he
described as occurring in great abundance in the so-called Cauda-
Galli Grit of the Ithaca group, Madison County. Vanuxem's fossil
was afterwards figured by Hall:|; under the name Spirophyton cauda-
galli Van. as occurring in North American strata referred to the
Devonian period, and described as a plant consisting of a slender
axis bearing a thin and broad spirally disposed "phyllome." Hall,
as well as several other authors, compared the fossil with certain
recent Algae. The authors of the Palaeobotanical volume of Zittel's
Handhnchdcr PalcBontologie i include Spirophyton with other similar
forms in the group Alectorurideae ; this group comprises examples
ranging in age from Silurian to Tertiary strata. There can be little
doubt that the fossils referred to Schimper's genus Alectoriorus || of
Silurian age, as well as those included in such genera as Cancello-
phycits 'i and Taonwrus "* from Mesozoic strata, without mentioning"
others from Upper Palaeozoic it and Tertiary rocks, might equally
well be designated Spirophyton. It is evident that Vanuxem's genus,
whatever its nature may be, is met with in sedimentary rocks of
almost all geological formations. Nathorst :|:.| speaks of the Alecto-
rurideae as occurring in such a position in sedimentary strata as must

* Annual Reports (97), pp. 19, -50, 04; (98), pp. 10, 49, 61 ; (99), p. 35.
t Vanuxem (42), pp. 128, 177, figs. p. 128. \ Hall (63).

§ Zittel (90), p. 54. |i Ihid., fig. 43, p. 55. "I Ihid., p. 57.

** Heer (76), pi. xlviii. ft Lesquereux (79), pi. A. \\ Nathorst (86=). p. 45.

104 Annals of the South African Museum.

lead to the conclusion that, if they be organic, they must have lived
on the spot where they occur as fossils. But as they are found in
deposits where they could have had no convenient support on which
to grow, the opinion to which he is driven is that S'piroijhyton and
allied forms are of mechanical origin and do not represent Algae or
other organisms.

Potonie" reproduces a figure of Spiropliyton cifeliense originally
published by Kayser t from Lower Devonian rocks ; he states that
by stirring sand in water he w^as able to produce an appearance like
that of the typical Sjjirophyton. In an exhaustive memoir byFuchs]:
Spiropliyton and other problematical forms are discussed in detail ;
this author describes an Eocene specimen in the Pisa Museum
having a diameter of 40 cm. ; he deals also with Tertiary examples
from the Flysch and elsewhere. Eeference is also made to the well-
known screw-like fossils from America i; — the Devil's corkscrew or
Daimonelia, and Spiraxis, described by Newberry, Barbour, James,
and others ; Fuchs expresses the opinion that these various fossils
are all of similar nature and do not represent impressions of plants.
He suggests that Spirophyton and similar markings may owe their
origin to some animal, probably a worm, which produced a spiral
cavity in sand or other sediment, and another comparison he makes
is with the spawn-bands (" Laichbander ") of certain nudibranchs
{e.g., Doris).

The examination of the Spirophyton specimens from South Africa
confirms my suspicion that this abundant " fossil " owes its origin to
mechanical causes. Spirophyton has so extended a geological range
that this fact alone makes one pause in describing it as a fossil alga ;
persistent types are well known to occur, but an organism with the
persistence of Spirophyton can only be recognised as such on con-
vincing evidence. The occurrence of this fossil in rocks, which are
in some instances almost destitute of undoubted organic remains, is
another fact to which some weight should be attached. But it is the
nature of the specimens themselves that furnishes the strongest
argument against the acceptance of Spiropliyton as an organism.
The strong and prominent ridges curving outward from a central
point do not convey the impression of records that are likely to have
been left by the thallus of an alga in a bed of sandstone ; the occur-
rence of the fossils in a vertical position traversing several inches of
a quartzose rock is a fact easier of explanation on the mechanical
theory of their origin. My view is that these Witteberg Spiro-

* rotoni6 (99), p. 39. f Kayser (72).

+ Fuchs (95), p. 410. S Newl)eiTy (8.')); James (95), &c.

Fossil Floras of Cajjc Colony. lOo

pliytons have probal)ly l:)een formed by swirling water -which has
left an impress in the form of obliquely striated and prominent
ridges (pi. xiv.) on the hardened sediment of an eddying stream
that flowed through the sand by a narrow vertical channel. I
ought to add that I have not made an extended examination of
Spirophytons from other localities, but as regards the South African
specimens my view is that they do not represent the remains of
plants.

Plate XIV., fig. 1 (57a).

The drawing shows the surface-view of a specimen, 13 cm. by

12 cm., in which the Spiropliyton characters occur on different
levels of the rock, but with no apparent regularity. In the centre
{a) there is an approximately cylindrical and obliquely placed cast,
from which the curved ridges and grooves radiate over the surface
of the stone, sloping slightly downward from the central point.
This axis may well be the cast of a central channel in which the
swirling water passed or flowed through the sand. As shown in
the figure, the surface of the rock is very uneven ; at c several
prominent ridges are seen covered by numerous oblique secondary
ridges. Similar obliquely striated and prominent ribs are shown
also at i at a higher level. There is no trace of any carbonaceous
matter. The strongly marked and prominent ridges form a striking
feature, and cannot, I believe, be explained on the assumption that
they have been formed by the impress of a leaf-like organ.

Plate XIV., fig. 2 (12a).

View of a specimen, 12 cm. broad, showing a portion of a Spiro-
pliyton impression. The centre from which the curving ridges
spring was probably close to the left-hand end of the drawing,
where the surface of the rock stands at a slightly higher level. The
edge of the impression (right-hand side of the figure) is fairly well
defined, and forms a definite boundary between the Spiropliyton
and the rock matrix. The surface is undulating, with occasional
broad curved folds which are traversed by numerous parallel ridges
less prominent than those shown in fig. 1. At d a few parallel ridges
are seen following a course oblique to the general trend of the
narrow ribs and grooves.

Other specimens : 1. A specimen from i mile South of Ladismith,

13 cm. by 13 cm., showing a truncated prominence in the centre,

106 Annah of the South African MuseiLm.

from which irregular ridges and grooves extend in double curves over
the face of the rock, forming a figure reminding one of the lines of
sparks given off from a rotating Catherine Wheel ; surface features
more obscure than in the figured specimens. The depth of the
grooves and the prominence of the ridges are in themselves a
weighty argument against the view that we have before us the
impression of an algal thallus.

3. From Worcester ; a small specimen similar to that shown in
fig. 2 ; 403.

(107)

BIBIilOGEAPHY.

AlIALITZKY, M. V.

(97) Note sur les nouvelles trouviiilles paleontologiques faites dans les depots
sablo-niarneaux peimiens de la Soukhona er de la Petite Dwina. Trav.
Soc. Nat. St. Petershurt], vol. xxviii.

(01) Sur la decouverte dans les depots permiens superieurs du nord de la
Russia, d'une flore Glossopterienne et de reptiles Parciasaiinis ct
Dicynodon. Coinpt. Rend. 1901, p. 591.

Annual Eeport of the Geological Commission, Cape Town.
(97) (98) (99) (00).

Andr.\e, K. T.

(53) Fossile flora Siebenbiirgens und des Banates. Ahli. k.-k. gcol. Eeichs.
vol. ii. Abth. 3, No. 4, 1853.

Arber, E. a. N.

(02) On the Clarke Collection of Fossil Plants from New South Wales.
Quart. Joiini. Geol. Soc, vol. Iviii., p. 1.

Blanford, H. F.

(75) On the age and correlations of the Plant-bearing series of India and
the former existence of an Indo-Oceanic Continent. Quart. Journ. Geol.
Soc, vol. xxxi., p. 519.

BODENBENDER, W.

(96) Beobachtungen liber Devon- und Gondwana — Schichten in der Argen-
tinischen Eepublik. Zeit. deiitscli. gcol. Ges., vol. xlviii., p. 743.

(02) Contribucion al conocimiento de la Precordillera de San Juan, de
Mendoza. Bol. Acad. Nat. Ciencias, vol. xvii.

BORNEMANN, J. G.

(56) Ueber organische Keste der Letten-Kohlengruppe Thiiringens. Leijizig,
1856.

Brongniart, a.

(28) Prodrome d'une histoire des vegetaux fossiles. Paris, 1828.
(28-) Histoire des vegetaux fossiles. Paris, 1828.
(45) See Murchison, R. I.

(49) Tableau des genres de vegetaux fossiles. (E.vtrait du Dictionnaire
d'liistoire naturelle, vol. xiii.) Paris, 1849.

Bronn, H. G.

(58) Beitriige zur triasischen Fauna und Flora der bituminosen Schiefer
Von Eaibl. Neues Jahrh. Min., 1858, p. 129.

BuNBURY, Sir C. J. F.

(61) Notes on a collection of Fossil Plants from Nagpur. Central India.
Quart. Joiini. Geol. Sue, vol. xvii., p. 325.

108 Annals of the South African Miiscum.

Caeeuthees, W.

(66) On some Fossil Coniferous Fruits. Geol. Mag., vol. iii., p. 534.

(68) British Fossil Pandanaceffi. Geol. Mag., vol. v., p. 158.

(69) On the plant remains from the Brazilian Coal beds, with remarks
on the genus Flcmingites. Geol. Mag., vol. vi., p. 5.

(69-) On some Undescribed Coniferous Fruits from the Secondary Rocks
of Britain. Ibid., p. 1.

(70) On Fossil Cycadean Stems from the Secondary Rocks of Britain. Trans.
Linn. Soc, vol. xxvi., p. 675.

(72) Notes on Fossil Plants from Queensland, Australia. Quart. Journ. Geol.
Soc., vol. xxviii., p. 1.

Cheist, H.

(97) Die Farnkriiuter der Erde. Jena, 1897.

CORSTOEPHINE, G. S.

See Annual Report, Geol. Commission, Cape Town (00).

Ceepin, F.

(75) Note sur le Pccopteris odontopteroides Morris. Bull. Acad. roy. Belg.,
vol. xxxix. [ii.] .

CUEEAN, J. M.

(85) On some Fossil Plants from Dubbo, N.S. Wales. Proc. Linn. Soc. N.S.
fVales, vol. ix., p. 250.

Dawson, Sir J. W.

(85) On the Mesozoic Floras of the Rocky Mountain Region of Canada.
Trans. B. Soc. Canada, sect, iv., 1885.

Deapeb, D.

(97) Notes on the occurrence of Sigillaria, Glossopteris, and other plant-
remains in the Triassic rocks of South Africa. Quart. Journ. Geol. Soc,
vol. liii., p. 310.

DUNKER, W.

(46) Monographie der norddeutschen Wealdenbildung. Braunschweig, 1846.

DUPONT, E.

(78) Sur la decouverte d'ossements d'Iguanodon, de poissons et de v^getaux
dans la fosse Sainte-Barbe du Charbonnage de Bernissart. Bull. Ac. B.
Belg., vol. xlvi. [ii.], 1S78, p. 387.

d'Eichwald, E.

(55) Lethiea Rossica (Anc. Periode). Stuttgart, 1855.

Etheridge, R.

(90) Note on the structure of Annularia australis Feist. Proc. Linn. Soc.

N.S. Wales, vol. v., pt. i., p. 47.
(90=) Lepidodcndron australc, McCoy. Its synonyms and Range in Eastern

Australia. Bcc. Geol. Surv. N.S. Wales, vol. ii., pt. iii., p. 119.
(92) ,SV(' Jack and Etheridge (92).

Ettingshausen, C. von.

(52) Beitrag zur niiheren Kenntniss der Flora der Wealden Periode. AhJi.

k.-k. geol. Beichs., vol. i., Ahth. iii., No. 2, 1852, p. 1.
(52-) Begriindung einiger neuen oder nicht genau bekannten Arten der Lias-

und der Oolithtlora. Ahh. k.-k. geol. Reichs., vol. i., Ahth 3.

Feistmantel, 0.

(75) Ueber des vorkommen von N(rggerathia foliosa Stbg. in dem Stein-
kohlengebirge von Ober Schlesien. . . Zeit. deutsch geol. Ges., 1875, p. 70.

(76) Fossil Flora of the Gondwana System. Mem. Geol. Surv. India, vol. ii.,
pt. i., Jurassic Flora of Kach.

Fossil Floras of Cape Colony. 109

76-) Notes on the age of some Fossil Floras in India. Bee. Geol. Surv.

India, vol. ix.
(77) Fossil Flora of the Gondwana System. ^Ictn. Geol. Surv. India, vol. i.,

pt. 2, Jurassic (Liassic) Flora of the Rajmahal Group in the Eajmahal

Hills.
(77-) Ihid., vol i., pt. 8 : Jurassic Flora of the Rajmahal Group from Golapili

(near EUore), S. Godaveri District.
(773) Ibid., vol. ii., pt. 2: Flora of the Jabalpur Group.

(79) Ibid., vol. i., pt. 4 : Outliers of the Madras Coast.

(79-) Ibid., vol. iii., pt. 1 : The Flora of the Talchir-Karharbari Beds.

(80) Ibid., vol. iii., pt. 2 : The Flora of the Damuda-Panchet Divisions.

(81) Ibid., vol. iii., pt. 1 : The Flora of the Talchir-Karharbari Beds.
(Supplement.)

(81=) 7///(/., vol. iii., pt. 3: The Flora of the Damuda-Panchet Divisions
(Pt. 2).

(89) Uebersichtliche Darstellung der Geologisch-Palseontologischen Verhiilt-
nisse Sud-Afrikas. Th. i. Abh. K. liohm. Gen. Wisa. Frarjue [vii.],
vol. iii.

(90) Geological and Palceontological Pielations of the Coal and Plant-bearing
Beds of Palaeozoic and Mesozoic Age in Eastern Australia and
Tasmania. Mem. Geol. Snrr. N.S. U'alex, Falieontoloijij, No. ii, 1890.

FiTTON, W. H.

(36) Observations on some of the Strata between the Chalk and the Oxford
Oolite in the South-East of England. Trans. Geol. Soc. [ii.], vol. iv.,
p. 103.

Fliche, P.

(00) Contribution a la tlore fossile de la Haute-Marne. Dnll. Hoc. Sri. Xa)ictj,
1900.

Fontaine, W. M.

(83) Contributions to the knowledge of the Older Mesozoic Flora of Virginia.

V. S. Geol. SuriK Mem. vol. vii.
(89) The Potomac or Younger Mesozoic Flora. U.S. Geol. Sure. Moii.,

vol. XV. Wasliinfitoi), 1889.
(92) Description of some Fossil Plants from the Great Falls Coal-field of

Montana.

Fontaine, W. M., and I. C. White.

(80) The Permian or Upper Carboniferous Flora of West Virginia and
South-West Pennsylvania. Second Geol. Snrv. Poin^^., 1880.

Fric, a., and E. Bayek.

(01) Studien im Gebiete der Bohmischen Kreideformation. Arcliir. Nat.
Landef:. von Bohnien., vol. xi.

Fkitsch, K. v.

(97) Planzenreste aus Thiiringer Culm-Dachschiefer. Zeit. Naturwis.-;., vol.
Ixx., p. 79.

FucHs, T.

(95) Studien iiber Pucoiden und Hieroglyphen. Denksch. Ak. ir/s.s-. IVicii.
vol. Ixii., p. 369.

Gaednee, J. S.

(86) On Mesozoic Angiosperms. Geol. Ma;i., vol. iii. [3], p. 193.

Geinitz, H. B.

(76) Ueber rhstische Pflanzen und Thierreste in den argentinischen
Provinzen LaRioja, San Juan, und Mendoza. Palccontographiea, Sajjpl.,
iii., 1879.

]10 Annals of the SoiUli African Museum.

Geylek, H. T.

(77) Ueber fossile Pflanzen aus der Jui-aformation Japans. PalcsontograpJnca,
vol. xxiv., p. 221.

GOEPPERT, H. R.

(86) Die fossilen Farrnkrauter. Nova Acta Ac. Cces. Leap. -Car. vol., xvii.,

supp., 1836.
(41) Ueber die fossile Flora der Quadersandstein Formation in Schlesien.

Nova Acta. Acad. Cas.-Leop., vol. xix., p. 97.

Gkev, G.

(71) Remarks on some specimens from South Africa. Quart. Journ. Geol.
Soc, vol. xxvii., p. 49.

Grigoriew, N.

(98) Sur la tlore paleozoique recueilie aux environs des villages Troitscoie
et Longanskoie dans le bassin du Donetz.

Hall, J.

(63) Contributions to Paleontology. IGth Aim. Rep. Univ. New York.

Haktz, N.

(96) Planteforsteninger fra Cap. Stewart i Ostgronland. Meddelelser Gronl.,
vol. xix., p. 217.

Hatch, F. H.

(98) A Geological Survey of the Witswatersrand and other districts in the
Southern Transvaal. Quart Jourii. Geol. Soc, vol. liv., p. 73.

Hadghton, S.

(60) On Cyclo^tifiDui, a new genus of fossil plants from the Old Red Sandstone
of Kiitorkan. Aninih and Maij. N((t. Hint. [3], vol. v., p. 443.

Heer, 0.

(75) Flora Fossilis Arctica, vol. iii., pt. ii.

(76) Flora Fossilis Helvetia. Zurich, 1876.

(77) Flora Fossilis Arctica, vol. iv., pt. ii.

(78) Flora Fossilis Arctica, vol. v., pt. ii.

(81) Contributions a la flore fossile du Portugal. Secc. Trab. Geol. Portugal.

(82) Flora Fossilis Arctica, vol. vi.

Heyer, F.

(84) Beitriige zur Kenntniss der Fame des Carbon und des Rothliegenden im
Saar-Rheingebiet. Bot. centralldatt, vol. xix., p. 248.

Holub, E., and M. Necmayr.

(82) Tiber einige fossilien aus der Uitenhage-Formation in Sud-Afrika.
iJcnhi^ch. Ahad. I['i,«., vol. xliv., p. 267.

Hosius, and von der Marck.

(80) Die Flora der Westfiilischen Kreideformation. Paheontoiirapldca, vol.
xxvi., p. 127.

Jack, R. L., and R. Etheridge, Jun.

(92) The Geology and Palaeontology of Queensland and New Guinea.
Urishaue, 1892.

Jaeger, G. F.

(27) tjber die Pflanzenversteinerungen welche in dem Bausandstein von
Stuttgart vorkommen. Stuttgart, 1827.

James, I. T.

(95) Remarks on Daiiitonelix, or Devil's Corkscrew, and allied fossils. Amer.
Geolofiist, vol. xv., p. 337.

Fossil Floras of Cape Colony. Ill

Jones, T. Rupert.

(84) On the Geology of South Africa. Nature, vol. xxx., p. -553.

Kayseb, E.

(72) Neue fossilien aus dem Eheinischen Devon. Zeit. dcutsch. Geol. Ges.,
vol. xxiv., p. 691.

KiDSTOX, E.

(89) Additional notes on some British Carboniferous Ljcopods. Annals Marj.
Nat. Hist., 1889, p. 60.

KOSMOVSKY, C.

(92) Quelques mots sur les couches a v6getaux fossiles dans la Eussie
orientale et en Siberia. Bull. Soc. imp. vat. Moscou., vol. v., p. 170.

Kkasser, F.

(00) Die von W. A. Obrutschew in China und Central Asien 1898-94 gesammel-
ten fossilen Pflanzen. Denksch. K. Akad. Wiss. ]Vicn, vol. Ixx.

KUETZ, F.

(94) Contribuciones a la palsophytologia Argentina. Pievist. 2Iiiit. de la Plate,

vol. vi.
(03) Remarks upon Mr. E. A. Newell Arber's Communication. Quart. Journ.
Geol. Soc, vol. lix., p. 25. '

Lapparemt, a. de.

(00) Traite de Geologie. Parh, 1900.

LiNDLEY, J., and W. Huttox.

(31-37) The Fossil Flora of Great Britain.

Lesquereux, L.

(79) Atlas to the Coal Flora of Pennsylvania. Second Geol. Surv. of Penii-
sylrania.

Lycett, J.

(79) A monograph of the British fossil Trigoniffi. Palccontographical Soc,
1872-79.

Mackie, S. J.

(62) The " Dragon Tree " of the Kentish Rag. Geologist, vol. v., 1802, p. 401.

Maxtell, G.

(33) The Geology of the S.E. of England. London, 1833.

McClelland, J.

(50) Report of the Geological Survey of India (1848-9).

McCoy, F.

(47) On the Fossil Botany and Zoology of the Rocks associated with the

Coal of Australia. Ann. and Mag. Nat. Hist., vol. xx., p. 145.
(74) Prodromus of the Palaeontology of Victoria. Dec. 1-6. Melbourne and

London, 1874-79.

Medlicott, H. B., and W. T. Blanford.

(93) Manual of the Geology of India (2nd ed.). Calcutta, 1893.

Moller, H.

(02) Bidrag tiU Bornholms fossila Flora. Lunds Univ. Arsxkrift., vol. 38,
Afd. 2, No. 5.

Morris, J.

(45) In Count Strzelecki's "Physical Description of New South Wales, &c."

p. 245. London, 1845.
(54) A Catalogue of British Fossils (2nd ed.'i. London, 1845.

112 Annals of the South African Museum.

MURCHISON, E. I., E T>E VeKNEUIL, & DE CoMTE AlEX. DE KeYSEKLING.

(45) Geologie de la Eussie d'Europe. London and Paris, 1845.

Nathoest, a. G.

(78) Om Floi-an i Skiines Kolforande Bildningar. Floran vid Bjuf. Hiift 1,
Srerig. Geol. Unders, 1878.

(78^) Beiti'iige zur fossilen Flora Schwedens. Stuttcjurt, 1878.

(79) Ihid. Floran vid Bjuf. Hiift 2.

(81) On spar af nagra everteberade djur M. M. och deras paleontologiska
betydelse. K. Srciifk. Vet. Akad. Hand., vol. xviii., No. 7.

(86) Om Floran i Skiines Kolforande Bildningar. Floran vid Bjuf. Sverig.
Geol. Undent., 1886.

(86^) Nouvelles observations sur des traces d'animaux. K. Srensk. Vetensk.-
Akad. Hand., vol. xxi.

(90) Beitriige zur Mesozoischen Flora Japans. DenkAchr. k. Ak. Wiss. Wieu,
vol. Ivii., p. 43.

(94) Zur fossilen Flora der Polarliinder. Th. i. Lief. 1. A'. Svensk. Vet.

Akad. Hand., vol. xxvi.
(97) Zur fossilen Flora der Polarliinder. Th. i. Lief. 2 : Zur Mesozoischen

Flora Spitzbergens. Ihid., vol. xxx.
(02) Beitriige zur Kenntniss einiger Mesozoischen Cycadophyten. Ihid.,

vol. xxxvi., No. 4.
(02=) Zur Fossilien Flora der Polarliinder. Th. i. Lief. 3. Ihid., vol. xxxvi..

No. 3.

Neujiaye, M.

(87) Erdgeschichte, vol. ii. Leipzig, 1887.

(82) See Holub and Neumayr.

Newbeeey, J. S.

(85) Descriptions of some peculiar screw-like fossils from the Chemung rocks.
Ann. N. York Acad. Nat. Sci., III., p. 217.

(88) Ehsetic Plants from Honduras. Amer. Journ. Sci., vol. xxxvi., p. 342.

(91) The Flora of the Great Falls Coal-field, Montana. Amer. Journ. Sci.,
vol. xli., p. 191.

Newton, E. T., and J. J. H. Teall.

(97) Notes on a Collection of Eocks and Fossils from Franz Josef Land, made
by the Jackson-Harmsworth expedition during 1894-96. Quart. Journ.
Geol. Soc, vol. liii., p. 477.

Oldham, T., and J. Moeeis.

(63) Fossil Flora of the Gondwana System. Vol. i., pt. 1 : Fossil Flora of the
Eajmahal Series in the Eajmahal Hills. Mem. Geol. Surv. India, ser. ii.
Calcutta, 1880.

POTONIE, H.

(93) Die Flora der Eothliegenden von Thiiringen. Ahh. k. 2J''f"SS. Geol.
Ijandes., vol. ix.

(99) Lehrbuch der PflanzenpalfEontologie. Berlin, 1899.

(00) Fossile PHanzen aus Deutsch- und Portugiescisch Ostafrika. Deutsch-
Oatafrika, vol vii. Berlin, 1900.

Eaciboeski, M.

(91) Ueber die Osmundaceen und Schizaeaceen der Juraformation. Engler'.s-
Bot. Jahrh., vol. xiii., p. 1.

(94) Flora Kopalna ogniotrwalych Glinek Krakowskich. Pamiet. Mat. Jirzy.
Akad. UiiiicjetnoTi.

Eatte, F.

(88) Notes on some Australian Fossils. Proc. Linn. Soc, N.S. Wales, vol. ii.,

p. 137.
(88^) Additional evidence on Fossil Salisburia; from Australia. Ihid., vol. i.,
p. 1078.

Fossil Floras of Cape Colony. 113

Renault, B.

(81-8-5) Cours de botanique fossile. Paris, 1881-85.

EOEMER, F. A.

(39) Die Versteinetungen cles norcldeutschen Oolilhen-Gcbirges. Ein Naehtrag.
Hanover, 1839.

Rogers, A. W., and E. H. L. Schwarz.

(01) Report on parts of the Uitenhage and Port Elizabeth divisions. Ann Rep.
Geol. Coinm., 1909, p. 3.

Sandrerger, F. v.

(90) Ueber SteinkohlenFormation and llothlieges des im Schwarzwald und
deren Flora. Jahrli. k. l<\ tjeol. lieichs., vol. xl., Heft i., p. 77.

Saporta, Le Marquis de.

(73) Paleontologie Francaise, Her. ii. : Vegetaux. Plantes .Turassiques,

vol. i. : Algues, Equisetacees, Charaeees, Fougeres. Pur'u, 1873.
(75) IJiid., vol. ii. : Cyeadees.
(84) Ihid., vol. iii. : Coniferes.

(91) Ihid., vol. iv. : Types proangiospermiques, et supplement final.

(93) Revue des travaux de paleontologie vegetale parus en France dans le cours
des annees 1889-92. Per. Gen. Rot., vol. v., 1893.

(94) Flore fossile du Portugal. Direct, trav. g^ol. Portugal, Lisbon, 1894.

SCHENK, A.

(67) Die fossile Flora der Grenzschichten der Keupers and Lias Frankens.

IVie.'ihaden, 1867.
(71) Die fossilen Ptlanzen der Wernsdorfer Schichten in der Nordkarpathen.

Pahrontofiruphica , vol. xix., 1871, p. 1.
(75) Zur Flora der nordwestdeutschen Wealdenformation. J'iil,<-onloiiraphica,

vol. xxiii., 1875-6, p. 157.
(83) Pflanzliche Versteinerungen (Richthofen's " China," vol. iv.).
(87) Fossile Ptianzen aus der Albourskette. Rihl. hot. Vlthronn and Hicnlcin,

Heft vi., 1887.

SCHIMPER, W. P.

(69-74) Traite de paleontologie vegetale.

Schimper, W. p., and A. Mougeot.

(44) Monographie des plantes fossiles du gr^s bigarre des Vosges. LeiprAci,
1844.

ScniMPER, W. P., and A. Schenk.
(90) Sec Zittel.

Schimper, W. P., and .7. Kochlin-Schlumberger.

(62) Les vegetaux fossiles du terrain de transition des Vosges. Strasshnrg, 1862.

SCHMALHAUSEN, J.

(79) Beitriige zur Jura-Flora Russlands. 3Iem. Acad. Imp. Sci. St. Pctersh.

[vii.] , vol. xxvii. , No. 4.
(87) Die Pflanzenreste der Artinskischen und Permischen Ablagerungen im

osten des Europaischen Russlands. jSIcdi. Com. Geol., vol. ii., No. 4.
(94) Ueber Devonisehe Ptianzen aus dem Donetz-Becken. Mem. Com. Geol.,

vol. viii., No. 3.

SCHWARZ, E. H. L.

(98) See Annual Reports (98), p. 51.

Sell ARCS, E. H.

(00) A new genus of Ferns from the Permian of Kansas. Kansas Univ. Quart.,
vol. ix., p. 179.

8

114 Annals of tJie South African Musejon.

Sewaep, a. C.

(94) Catalogue of the Mesozoic Plants in the Department of Geology, British
Museum (Nat. Hist.). The Wealden Flora, pt. i.

(95) Catalogue of the Mesozoic Plants in the Department of Geology, British
Museum. The Wealden Flora, pt. ii.

(96) Notes on the Geological history of Monocotyledons. Annuls Bot., vol. x.,
p. 205.

(97) On the leaves of Ileiincttites. Froc. Phil. Hoc. Cmnhvithje, vol. ix., pt. v.,
p. 273.

(97-) On the Association of Siqillaria and Glo^soptcris in South Africa. Quart.
Joitrn. Geol. Soc, vol. lii'i. (1897), p. 315.

(98) Appendix to Dr. Hatch's paper — A Geological Survey of the Witswaters-
rand and other districts in the Southern Transvaal. Quart. Juurn. Geol.
Soc, vol. liv., p. 92.

(98) Fossil Plants for Students of Botany and Geology, vol. i. Caiiihridjie,
1898.

(00) Catalogue of the Mesozoic Plants in the Department of Geology, British
Museum. Jurassic Flora I.

(00=) Notes on some Jurassic Plants in the Manchester Museum. Meiit. (utd
Froc. Manclu'ster Lit. and Fliil. Soc, vol. xiv., pt. 3, p. 1.

(003) La Flore Wealdienne de BernLssart. Mihit. Mua. d'liixt vat. Bclf/.,
vol. i.

(02) Article Paleobotany (Mesozoic), Encijclopcedia Brittanica, vol. xxxi.,
p. 421, 1902.

Seward, A. C, and E. Dalk.

(01) On the structure and affinities of Dipteris, with notes on the geological
history of the DipteridinsB. Fhil. Tran^. R. Soc, vol. 194., p. 487.

Seward, A. C, and S. 0. Ford.

(03) The Anatomy of Todca, with notes on the Geological history and
affinities of the Osmundacea3. Tran:^. Liirn. Soc, vol. vi., p. 237.

Seward, A. C, and J. Gowan.

(00) The Maidenhair Tree {Ginla/o hiloha L.). Ann. of Botany, vol. xiv., p. 109.

Sharpe, D.

(5G) Description of fossils from the Secondary rocks of Sunday Kiver and
Zwartkop River, South Africa. Trans. Geol. Soc, vol. vii. (ii.), p. 193.

Shirley, J.

(98) Additions to the Fossil Flora of Queensland. Dep. Mines ; Geol. Sitrv.
Bull., No. 7. Brisbane, 1898.

Solms-Laubach, Graf zd.

(91) Fossil Botany (English translation). O.rford, 1891.

Solms-Ladbach, Graf zxt, and G. Steinmann.

(99) Beitriige zur Geologic und Palseontologie von Siidamerika. Neues Jalrrh.
Min. Beilaiiehand, xii., 1899, p. 593.

Stokes and Webb.

(24) Descriptions of some Fossil Vegetables of the Tilgate Forest in Sussex.
(The authors' names do not appear in the title of this paper.) Trans.
Geol. Soc. [ii.]', vol. i., 1824, p. 421.

Stow, G. W.

(71) On some points of South African Geology. Qiiarl. Jourii. Geol. Soc,
vol. xxvii., p. 28.

Stur, D.

(75) Die Culm Flora. Abh. k,-l\ Geol. Beiehs., vol. viii,

Fossil Floras of Cape Colony. 115

SZAJNOCHA, L.

(88) Uebcr fossile Pflanzenreste ann Cachcuta in den Argentinischen Eepublik.

Sitzb. k. Ahtd. Wins. Wicii, vol. xcvii. Ahth. 1, p. '2r.».
(91) Uber oini^c Carbone Pfianzanicste aus der AiRcntinischon Itcpublik.

Sith. k. A bid. I(';.s.s'. Wlen, vol. c. .[hth. 1, p. 203.

Tate, E.

(67) On the Secondary Fossils from South Africa. Quurl. Jotirit. Geol. Soc,
vol. xxiii., 1867, p. 139.

Tenison-Woods, J. E.

(83) On the Fossil Flora of the Coal Deposits of Australia. Proc. Linn. Soc,
N.S. IVales, vol. viii., pt. 1, p. 37.

Unger, F.

(64) Eeise der Oesterreichischen Fregatte Novara urn die Erde. Geologischer
Theil, vol. i., Abth. ii., p. 1. Wicu, 1864.

Vanuxem, L.

(42) Natural History of New York. Part III.

Velenovsky, J.

(85) Die Gymnospermera der Bohmischen Kreideformation. Prag., 1885.

(88) Die Fame der Bohmischen Kreideformation. Ahh. k. hohm. Ges. Wii^s.
[vii.], vol. ii., 1888.

Ward, L. F.

(95) The Potomac Formation. Ann. Rep. U.S. Geol. Sitrv., 1895.

(96) Some Analogies in the Lower Cretaceous of Europe and America. Ann.
Pep. U.S. Geol. Sure, 1896.

(99) The Cretaceous Formation of the Black Hills as indicated by fossil plants.

U.S. Geol. Surv., 1899.
(00) Status of the older Mesozoic Floras of the United States. U.S. Geol.

Siirr., 1900.

Weiss, E.

(69) Fossile Flora der jiingsten Steinkohlenformation und des Eothliegenden
im Saar-Eheim Gebiete. Bonn., 1869-72.

White, C. D.

(89) Carboniferous glaciation in the Southern and Eastern Hemispheres, with
notes on the Glossopteris Flora. Amer. Geol., vol. iii., p. 299.

Yokoyama, M.

(89) Jurassic Plants from Kaga, Hida, and Echizen. Journ. Coll. Sci. Japan,
vol. iii.

(94) Mesozoic Plants from Kozuke, Kii, Awa, and Toza. Ihid., vol. vii.,
pt. 3.

Zeiller, E.

(82) Examen de le flore fossile des Couches de Charbon du Tong-King. Ann.
Minex, vol. ii. [viii.] , p. 299.

(83) Note sur la fiore der bassin houiller de Tete. Ann. Mines, vol. iv. [viii.],
p. 594.

(95) Note sur la flore fossile des gisements honillers de Eio Grande do sul.
P>ull. Soc. ijeol. France, vol. xxiii., p. 601.

(96) Etude sur quelque Plantes fossiles . . . de Environs de Johannesburg.
Soc. (jeol. France [iii.], vol. xxiv., p. 349.

il6 Annals of the South African Museum.

(96^) Remarques sur la flore fossile de I'Altai. Bull. Sac. geol. France,
vol. xxiv. [3], p. 46G.

(97) Les Provinces botaniques de la fin des Temps primaries. Rev. Gen. Sci.,
1897.

(98) Sur la decouverte, par M. Amalitzky, de Glossopteris dans le Termien
Superieur de Russie. Bull. Soc. hot. France, vol. xlv., p. 392.

(00) Klements de Paleobotanique. ParU, 1900.

(0"2) Observations sm- quelques plantes fossiles des Lower Gondwanas.

Fa Upontoloqia Indica .
(02=) Flore fossile des Gites de Charbon du Tonkin (Atlas). Pari.% 1902.

ZlONO, A. DE.

(56) Flora fossilis formationis Oolithicai, vols i. and ii., Paduva, 1856.

ZiTTEL, K. A.

(90) Handhuch der Palseontologie. Abth. ii., Schimper, W. P., and A. Sehenk.
Mnnkh and Leipzig, 1890-1.

( 117

EXPLANATION OF PLATES.

Plate I.

0»)?/c/n'()j).s';.s- iiunitclli (Brongn.).

Fig. 1. Natural size (4'2.Sc). White River, Dunbrodic (Goclhoutboom)

,, 2. X 1^ (391c).

,, 3. X 2 (407c).

„ 4. X 2 (401c).

Plate II.

Cladtijililcliis brtiinmnid (Dunk.).

Fig. 1. Natural size (843c). Near mouth of Bezuidenhout's Pi.
hi. X 4 ,, ,,

2. Natural size (344c).
2rt. X 4

3. X 3 (364c).

4. Natural size (342c).
4«. X 3
4/;. x4
6. X U (361c).

TirnioptcrU sp.

Fig. 5. Natural size (172d). Two miles east of Herbertsdale
„ ort. X 1^ (193d). „ „

Sphouipteris fittoni Sew.

Fig. 7. Natural size (249c). Near mouth of Bezuidenhout's K.

„ la. X 3

„ V>. x2

„ H. X 2 (376c).

,, 8a. X 4 ,,

118 Annals of the South African Museuvi.

Plate III.

Zdinitcs recta (Tate).

Fig. 1. Natural size (257c). Near mouth of Bezuidenhout's E.
„ 1(7. X 2 (258c).
„ 2. Natural size (288c).
M 3. „ (A).

Plate IV.

Nilxsonia tatci sp. nov.

Fi". 1. )

" ,^ f Natural size (291c). Near mouth of Bezuidenhout's B.
)> -'• '

Cycadolepis jcnkinaiana (Tate).

Fig. 3. Natural size (315c). Near mouth of Bezuidenhout's B.
„ 4. ,, (280c).

„ 5. x3 (319c). ,, „

,, 6. Natural size (311c). ,, ,,

Plate V.
Onycliiopsix maiitclli (Brongn.).

Fig. 1. Natural size.
(Type-specimen of Tate ; Museum of Geological Society of London, No. 11,114.

Bentitedtia sp.
Fig. 2. Natural size (292c). Near mouth of Bezuidenhout's B.

Zamites rtihulcici, (Tate).

Fig. 3. Natural size.
(Type-specimen of Tate ; Geol. Soc. Museum, No. 11,109.)

Zamites viorrhii (Tate).
Fig. 4. Natural size.
(Type-specimen of Tate ; Geol. Soc. Museum, No. 11,108.)

Zainiteti afrieana (Tate).

Fig. 5. Natural size.
(Type-specimen of Tate; Geol. Soc. Museum, No. 11,110.)

Fossil Floras of Cape Colony. 119

Plate VI.

Coniten sp. a.

Fig. 1. Niitunil size (2530). Near mouth of Uezuideniiout'.s \\.
,, Irt. X 5 ,, ,, ,,

,, 3. Natural size ('298c). ,, „

Conitea sp. j^.

Fig. 2. Natural size (310c). Near mouth of Bezuidenhout's l\.
„ 2"- X 2i

" 2 " i> I)

Arciticarites w<jc'i:'<i sp. nov.

Fig. 4. Natural size (29.Sc). Near mouth of Bezuidenhout's E.
,, 0. ,, „ ,, ,,

„ 6. „ (295c).

,, 7. „ (300c).

Ziimite.'> recta (Tate).

Fig. 8. Natural size (298c). Near mouth of Bezuidenhout's i!
„ it. „ (32SC).

,, 10. Considerably enlarged (a). ,, „

,, 11. X Ih (342c). „ „

,, 12. Natural size (280c). ,, „

IJraclii/piiijUiiin sp.

Fig. 13. Natural size (299c). Near mouth of Bezuidenhout's 11.
,, 18. X 3 (348c). „ ,,

Pliuttd iiicertw aedis.
Fig. 14. Natural size (332c). Near mouth of Bezuidenhout's E.

Taxites sp.
Fig. 15. Natural size (182d). Herbertsdale.

CladophlehiK denticidata (Brongn.) forma utherstonei.

Fig. 16. Natural size (170d). Herbertsdale.
„ 17. X U

120 Annals of the South African Museum.

Plate VII.

TJiiuufeldia odojitopteroides (Morr.).
Fig. 1. Natural size (21e). Kenigha Hiver.

Steiiojiteria cloiKjata (Carr.).

Fig. 2. Natural size (21e). Kenigha River.
„ 3. ,, (B). Stormberg.

„ 3«. „ (B).

,, 3<(. Ultimate segment enlarged.

Calliptcridiiim stoDnbergcnse sp. nov.
Fig. 4. Pinnules enlarged (x 2) (10). Kenigha River.

,, .'). Apex of frond (31e). ,,

Figs. 6, 6ff. Natural size (39e). „

Fig. 6a. Pinnule enlarged (x 2).

Thinnfeldia odo)tt(y2)teroides (Morr.).

Fig. 7. Natural size (477d). Maclear village.

Figs. 8, 8(1. Natural size (45e). Edward's Hope, Matatiele.

Plate VIII.

Thinnfeldia rhoinhoidalis Ett.
Fig. 1. and a fragment of Stenopteris elonguta (Carr.) (C). Stormberg.

Callipteridiiini atormhergcnse sp. nov.
Fig. 2. Natural size (46e). Edward's Hope, Matatiele.

Baiera storinJ)ergensis sp. nov.
Fig. 3. Natural size (x).

Tceniopteris carriithersi (Ten. -Woods).

Fig. 4. Natural size (1). Kenigha River.
,, 5. ,, (14e).

„ 0. ,, (497d).

Thinnfeldia odontopteroides (Morr.).

Fig. 7. Natural size (22c). Kenigha River.
„ 8. ,, (d). Cyphergat,

Fossil Floras of Cape Colonii. 121

Tl-ATK IX.

I'hiriiicojisis doiiiiiiid (Morr.).
Fi;^. 1. Natuial size (C). C.vplier^'ut.

St,icliiiiij)iti/.< s[). (cf. Sjilu'iiolcpi^ rli„tiii!. Oein.).
Fi.t^'s. -2. -111. Xaturiil size [Vl). KtiiiuliM 'Wwev.

Slnihililr^ sp.
Pis- '^. Niitmul size (i^M aiid I'.tod). Tina Diift.

< 'liiro/itcri^ riiiicdiii (Can'.).
l'"i^'. I. Natmal size (C). Cypheriiat.

S<-lii:<'iiciini 1,-noiscrt s[). iio\-.
Fio'. -). Natural size (:J0S(1). Matatiele villa<,'P.
-> •■'• ., (4«7(1). Tina Drift.

'I'liiinij'clilid iKlinilKjilcrdiilcs (Morr.).
Figs. 7, 8. .Mai^nitied rra^iiientH of eutiele showiii^i stoniata, itc.

I'hiriiii-iiji^i^ cliiiiijntn (Morr.).

Fig. 9. Natural size (d). Cyphergat.
,, 10. ,, ((;<)). ^folteno.

ri.ATi: X.

yciiroptcrldhuii niliduiii Feist.

FTgs. 1, 1<(, Ih. Pinnate frond (XXII.). Vereeniging.
Fig. 1((. Pinnule approximately natural size ((/ in fig. 1).

„' 1/'. ., ,.' ,. (/' ,, ).

.. '2. Basal portion of frond (XX\T.). Vereeniging.

<!l<>.<s(ijilcr(s liroiriiiiiiKi BrouKii. var. iiiiii/tsti/nHii.
Fig. :!. Natural size (XXII.). Vereeniging.

(ilo.tsiiplc'ris hnnriii'iii'i var. iiidlcd.
Fig. 4. Natural size (XXII ). Vereeniging.

So'ijiic ralli h iji'i ( •■-■ Ji isloji i.
Fig. •'5. Natural size (XX.). Vereeniging.

122 AmiaJs of the South African Museum.

Pi-ate XI.

Jloth rode nd ran ledii sp. nov.

Fig. 1. Natural size (XXL). Vereeniging.
,, la. Scar slightly enlarged.
,, Ih. Portion h of fig. 1 slightly enlarged.
,, 4. Natural size (B).
Figs. 5 and (>. Natural size (XX.). Vereeniging

Tliinnfeldiu odntiKiptcroidcs (Morr.).
Fig. 2. Natural size (44e). Edward's Hope, Matatiele.

Stciii)2>tcris cloiHidlii (Carr.).
Fig. 8. Natural size ((>). Kenigha River.

Pi.ATj; XII.

l'si/iiiii(ii>lii/llitiii kid>it(iiil sp. nov.
Fig. 1. Natural size (XXIII.). Vereeniging.

fllossoptcrin?
,, 2. ?P>hizonie and roots (XX.). ,,

Plate XIII.

(llnx.-ioptcrix hnuniidiiii var. iiidiru.
Fig. 1. Secondary veins (x 2) (Z). Worcester Station.

Nirjiiicrdtliidjisi^ hi^Iiijii (Bunh.).

Fig. 2. Natural size.
,, 3. Imperfect leaf, natural size (XXV.). Vereeniging.
,. 4. ,. „ (XXIV.).

(uiii(j({iiii<}itn-h (■jicldiitcroidc^ Feist.
Fig. •>. Natural size (A). Vereeniging.

Pi.ATi: XIV.

Siiirdpln/ldii.

Fig. 1. Letta's Kraal, Worcester division.
,, 2. Zout Kloof, 3 miles F. of Ladismith.

Arm S Afr Mus. Vol 4

PI

/ y-*'

- ^ \

^ f

/ /

f ./ ^

i i

/ ^ "\

'^-^ ;', '. /v

/'\

»'V4.^ 'ii:i

fj

/

Onychiopsis.

4

SlVv^lson^ Cambridge .

Ann.S.Afr. Mus.Vol.4

PI. 2,

->

i\/^,ii

'^^^M
^f'

1
I

x3

X I ^2 5a

x2 ' ^^ZP^""^

' ' \ '^

x4

8a

E- Wilson, Cambridge

Filicales.

Ann. S. Afr. Mus. Vol.4.

PI. 3

'"^

E }^'i 1 s on , C aiTiO ridge.

ZaiTLit e :

Ann.S. Afr. Mus.Voi.4.

P1.4.

Fig.l. "^P^^

-('i|):KiilM!|U, ^^iis;-^

■flj

x3

Fiq.5.

Fig. 2. [^'

'Y*f4

Fiq 3

Fiq.4

y

J

E.Wilson, Cambridge.

Cycadophjba

PI. 5.

E .Wilson, Cambridge .

Ann.S.Afr. Mus.Vol.4.

PI. 6.

16.

E .Wilson, Cambridge .

Coniferales etc.

Ann. S. Afr Mus.Vol.4

PLVll

3a

.N.^

cr

^"•-

'^'m^C-:^

^.-n

8a

k Uf ,vfk

^-"•y

7.

)'

^ V,

0

/

1 i

^\i\ ^

\

\

-,-;-3.^i

l**^-

.m,

<

^^h^r?^^^^

Filic ales

Ann. S. Afr. Mus. Vol.4,

Pl-VII'

■"^^u:^^

i^^Hi

L.

■ .1/

\

X. ' I .J

8 '■•

7 I

X.

f

■«:^ ■■ ■ y

t

E Wilson, Camfcriilgf

Gmkqoales , Filicale;

Ann. S. AtV. Mus. Vol.4.

PI. IX.

II

w

'. 1 i

h /

^/^lil)\\

^^

W'l '\

:^^

i!!'i

|.v.

Si

EVuiScii, Cambricke.

Equlsetalss etc.

Arm. S. Afr. Ivfus. Vol.4.

.^N

i^, Fi0. la.

^ Fig. IK

/

Fi^. 2.

'^^■.^^.~',< •

o' ..-r^

y./-"'

;s^v>\rx,

k

Neuron h

/

Fh^.1.

^N

Pl.X

Fi^.3.

/ ^^^

X'

Fig. 4.

■^^-

^^g-

E .Wilson, Cauibvidge.

Ann. S. Afr. Mus.Vo].4.
1. / ■-

PI. XL

E.WilsoiVj CaJntridcje.

Lvcopodiale s etc,

kun. S. Afr. Mils. Vol. 4.

\

\

^

/

1, N

\

\ t.

Fi^.i.

i

.t|>

J..

..-»=^

^j^f»-'-*«'^

Psyvn-f

%

^

^--.^-S^^

^^STi

£.V/ilacT., Catr.oriijt:

livllum.

Ann. S. Afr. Mus.Vol.4

PI. XIII,

1. X 2

4

in

ft '

'i h

/

^'/.

m i

'l;

jl

i

;' '

•I'

1)

I

1

' i i

1

1

3.-/5»vtf*'

\

u

.'J^

u^-

<^

<f^!s:

'■^■^■-'^m^iii^j.

""S^^^-

E WUson, Canlbridq

Noeggerathiopsis etc.

Ann. S.Afr. Mus.Vol.4.

PI. XIV

.^~^y

^''^ v^

:• *«hf rt-»^

■»c-

S, V.

' ^, '*^3«^^ :

'0 V^

Fi^ 2 j^/ .

\ 'J^

■, ''/

>/

E.Wilson, Camlsridge

Spirophy ton.

IT

G

A

Seward, A. C. /Fossil tlo

3 5185 000