Pinna atherstoni Sharpe and P. sharpei Tate with Jurassic forms, for it is evident that these representatives of the genus lack any strong distinctive characters which alone would render them of value in an estimate of geological age.

Mytilus baini Sharpe was compared with M. sowerbianus d'Orb., from the Oolites of Europe. It is a Modiola which certainly has characters of form and ornamentation closely similar to those which distinguished M. sowerbiana (d'Orb.) and M. perplicata (Etallon); but this type of Modiola is by no means confined to Jurassic rocks, having also a widely distributed representative in the Cretaceous M. flagellifera (Forbes), to mention a well-known example.

Pholadomya dominicalis Sharpe was thought to resemble Liassic and Oxfordian forms, but it may equally well be brought into comparison with Lower Cretaceous representatives; for example, the European Aptian shell ascribed, rightly or wrongly, to P. pedernalis Roem.*

Astarte pinchiniana Tate was brought into comparison with A. pumila Goldf. and A. excentrica Morr. and Lyc., from the Great Oolite of Europe. The African shell differs in important respects from these, and, as will be pointed out below, it probably belongs to the sub-generic group Eriphyla and is certainly most closely comparable with Cretaceous forms.

Berenicea antipodum Tate was believed by Tate to be related to B. striata Haime, from the Lower Lias of France, but has been regarded by Prof. J. W. Gregory † as a close ally of the Cretaceous B. gracilis (Milne Edwards).

Two Uitenhage Serpula were identified by Tate with S. filaria Goldf. and S. plicatilis Münst., from the Inferior Oolite of Europe. The specimens ascribed to S. plicatilis represent an adherent form sharply carinated on the back, with a weaker carinal angle on either side and a circular aperture, but the true S. plicatilis Goldf. has only a single central, longitudinal keel. S. quinquangularis Goldf. ‡ (Upper Jurassic) stands much closer in its characters, and a Serpula scarcely distinguishable from this occurs in the Lower Cretaceous of Europe, and may be found named S. quinquangularis in some collections. S. filaria Goldf. is a solitary, wholly adherent form, commencing with a coiled stage. So far as can be seen from the imperfect African specimen named S. filaria by Tate, this is a clustering form consisting of crowded, narrow cylindrical tubes, and * Pictet and Renevier (1), p. 60, pl. vi., fig. 7, 1855; Moesch (1), p. 93, pl. xxxiii., fig. 2. † Gregory (1), p. 112. Goldfuss (1), Band i., p. 230, pl. 68, fig. 8 (1831).

it is not distinguishable from the English Lower Cretaceous Serpula that have been variously referred to Serpula filiformis J. de C. Sow, and S. plexus J. de C. Sow. We thus see that in the case of the two Trigonia and the two Serpula which Tate identified with European Jurassic types, the determinations were so far erroneous that they are now bereft of the significance which was attached to them as evidence for the age of the fauna.

We may further examine the evidence for age afforded by some of the more conspicuous bivalves, and firstly I may briefly refer to the representatives of the genus Trigonia, which in themselves lend. such a characteristic aspect to the fauna. Although the testimony of these is perhaps not quite so conclusive as Lycett supposed, still, the balance of evidence furnished by members of this genus may be safely relied upon to give a fair indication of age, even when considered apart from associated forms, and the material we are dealing with in the present instance is fortunately of an exceptionally favourable character. In the first place there is Trigonia ventricosa, T. kraussi, and T. rogersi, all possessing the characters of the section Scabræ a division which, so far as we know, is elsewhere only represented in rocks of Cretaceous age. These might be considered to be balanced by the occurrence of T. tatei, a typical member of the section Costatæ, but although this division of the Trigonia has its maximum development in the Oolitic rocks, it is also represented in strata of Lower Cretaceous and even later age.

Trigonia peninsularis Coquand,* from the Aptian of Spain, certainly shows characters of degeneration not shared by T. tatei, but the Oomia strata of Cutch have yielded two normal representatives of the Costatæ, which, together with some degenerate derivatives of this section, are accompanied by Trigonia which exhibit a decidedly Cretaceous aspect. Another unmodified costate form, T. angustecostata Behr. occurs in the Argentine Republic in strata which are regarded by Behrendsen as Upper Cretaceous, while a typical member of this section has been described by E. Ascher from the Grodischter beds (Hauterivian) of Silesia. ‡

Trigonia herzogi (Goldf.), another very characteristic Uitenhage form, was believed by Lycett to belong to the Quadratæ, which, so far as known, are exclusively Cretaceous. Steinmann,§ however, pointed out that Trigonia herzogi differs in several important par

* Coquand (1), p. 129, pl. xxiii., fig. 3.

+ Behrendsen (1), p. 6, pl. 3, fig. 7.

Ascher (1), p. 159 [25], p. xiii. [ii.], fig. 10.

§ Steinmann (2).

ticulars from all known Quadratæ, and found that the special features which distinguish this Uitenhage shell from the Jurassic Clavellatæ on the one hand and the Cretaceous Quadrata on the other, are shared by Trigonia transitoria Steinm. from Lower Cretaceous rocks in Bolivia, Chili, and the Argentine Republic. He therefore suggested a new group-name, the Pseudo-quadratæ, to comprise these two Trigonia, which were the only examples known to exhibit these peculiar distinguishing characters. Trigonia neuquensis Burckhardt,* which occurs with T. transitoria in Neocomian rocks at Las Lajas on the Rio Agrio (Argentine), falls under the same heading, while T. mamillata from the Oomia strata in Cutch must also be referred to this group. Still another form, described in these pages (T. holubi), is clearly referable to the same category. Although the five members of the Pseudo-quadratæ known to us are all similarly differentiated from the typical Clavellatæ, it is not certain that they are all so closely related to one another as might at first sight appear to be the case; some, at least, may possibly have been independently evolved from clavellate ancestors. Nor is it by any means certain that they bear any close and direct relationship to the true Quadratæ, and though definite indications on this point fail us, it is quite probable that the Quadrata and Pseudo-quadratæ represent parallel and independent groups of derivatives arising from Jurassic Clavellatæ. It is therefore evident that the Trigonia of the group Pseudoquadratæ cannot yet be considered in themselves to furnish quite such reliable guidance to geological age as the true Quadratæ, but at the same time their general aspect, by comparison with the Quadratæ, strongly suggests a Cretaceous age, while the known South American representatives must, in fact, be regarded as Neocomian. Some of the most important features wherein the Quadratæ differ from the Clavellatæ are also exemplified in the Pseudo-quadratæ. Lycett laid great emphasis on the presence or absence of sculpture on the escutcheon in distinguishing between the Jurassic Clavellatæ on the one hand, and the Cretaceous Clavellatæ and Quadratæ on the other, and in regard to this feature all the Pseudo-quadratæ agree with the Quadratæ in having coarsely nodose ornamentation on the escutcheon. Further remarks on the points of agreement between these groups are appended to the description of Trigonia holubi.

Trigonia conocardiiformis (Krauss) was thought by Lycett to fall

Burckhardt (2), p. 74, Taf. xiv., figs. 4-6.

+ Kitchin (1), p. 100, pl. ix., figs. 8, 9; pl. x., figs. 1-3.

Remarks on this subject will be found to follow the description and the discussion of the relationships of Trigonia holubi, in these pages.

within the section Scabre, but a near relationship to members of this division is scarcely to be inferred from the adult characters displayed by this peculiar form, while a study of the young shell does not plainly reveal its affinities. A general comparison, however, shows that this Trigonia differs from the Jurassic Clavellatæ in a manner somewhat analogous to that which distinguishes the Scabræ, and a very closely similar and probably intimately related form occurs in the Neocomian of the Argentine Republic.

On the whole, so far as this Trigonia-assemblage is comparable with European forms, it must certainly be considered to display a Cretaceous rather than a Jurassic character, though when regarded collectively, it is without counterpart in the European area. The significance of other Trigonia included in this fauna will be discussed presently.

Still further indications of geological age are to be derived from a comparison of the Uitenhage lamellibranchs with European types. One of the shells submitted to me, which proves to be identical with a form previously described by Sharpe, who did not recognise its true generic position, is an example of the well-characterised genus Thetironia, which in Europe is widely distributed in Lower Cretaceous and higher strata, though it has not been recorded from rocks of an earlier age. A second species of Thetironia is also included in the collection. Solecurtus is another genus not known to occur in rocks older than Cretaceous, and a representative of this must now be added to the list of Uitenhage molluscs. The specimens sent to me for examination also include examples of Pecten which can only be identified with the Cretaceous P. orbicularis J. Sow. and P. cottaldinus d'Orb., while a third form may be most aptly brought into comparison with P. subacutus Lam. A representative of the Cretaceous genus Anthonya must also be recorded in this connection.

Gervillia dentata Krauss, though belonging to the group of G. aviculoides,* typically represented in Jurassic rocks (G. deecki Frech; G. aviculoides Sow.), has a close counterpart in G. anceps Desh. and G. sublanceolata d'Orb., in the Lower Cretaceous of Europe. A form closely similar, and perhaps identical, occurs also in the Neocomian of German East Africa. The long lateral tooth which Krauss thought to be so highly distinctive of G. dentata as almost to justify the establishment of a new genus, is a normal feature of the group, and is well developed in G. anceps. Another common

*

Frech (1).

† Frech (1), pp. 612, 613.

and characteristic Uitenhage shell is Exogyra imbricata Krauss, and this belongs to a general type well exemplified in the Lower Cretaceous strata of Europe and South America. Lima neglecta Tate, belongs to the sub-generic division Mantellum, and most closely resembles forms which occur in the Lower Cretaceous of Europe. Reference has already been made to Ptychomya complicata (Tate), which is a typical representative of a genus widely distributed in the Lower Cretaceous rocks, where it first makes its appearance.

It is abundantly evident, then, that many of the most important and distinctive molluscan forms, amongst those with which we are dealing, give similar indications of geological horizon. There can be no doubt, indeed, that the marine Uitenhage fauna is of Neocomian age, though it is perhaps not possible to arrive at a very precise and positive estimate of equivalence with the divisions of the Neocomian in Europe, owing to the want of close coincidence in the faunas as developed in such widely separated regions. But Holcostephanus atherstoni (Sharpe) and H. wilmana sp. nov., if not actually represented in Europe, have very close allies there in the Valanginian and Hauterivian, and the Uitenhage Marine Beds may, in my opinion, be certainly correlated with these divisions. Further, it seems to me highly probable that the Marine Beds represent not more than the strata at the top of the Valanginian and the base of the Hauterivian.

The possibility that the lowest marine strata may be of greater age than Neocomian is not only most remote, but is contradicted by the paleontological evidence, so far as this goes. Fossiliferous strata, yielding a marine molluscan fauna characterised by the prevalence of Gasteropoda and the remains of oysters, occur at Dunbrodie (Sunday's River) associated with plant-bearing beds, and are in the lower part of the Uitenhage Series; according to Messrs. Rogers and Schwarz they may be classed with the so-called "Wood Bed" series. Amongst the fossils collected here are Actæonina atherstoni (Sharpe), Cyprina rugulosa Sharpe, and a Pecten which I have identified as Pecten cottaldinus d'Orb. While Acteonina atherstoni is found also in the highest part of the Marine Beds on the Zwartkop's River and with the characteristic Trigonia on the Sunday's River, Stow recorded the occurrence of Cyprina rugulosa with similar associates. Although the Wood Bed series was not seen in the Zwartkop's River Valley below Uitenhage, the lowest part of the Marine Beds in the Zwartkop's River section, exposed in a clay-pit near Rawson Bridge, yielded Actæonina atherstoni and