"Fossils of Piltdown"
Wealden Mammalian Fossils
William A. Clemens
 Abstract. Only two of the previous identifications of specimens considered to be teeth of Wealden mammals can be accepted without reservation. A special collecting technique including both chemical and mechanical processes facilitated the discovery of eight more Wealden mammalian fossils. Five were found in the Cliff End Bone Bed, a part of the Ashdown Beds, and the remainder in the Paddockhurst Bone Bed, a part of the Grinstead Clay. These fossils give additional information about the morphology of the multituberculate Loxaulax valdensis and demonstrate the presence of a symmetrodont and eupantothere in England in the early (pre-Aptian) Cretaceous.
The fossil record of Mesozoic mammals is limited. Large collections of mammalian fossils of late Jurassic age have been made in the Purbeck Beds of England and the Morrison Formation of the United States. The Albian mammalian fauna is known from fossils found in the Trinity Sand of Texas. The only record of mammals that certainly lived in the interim between the Purbeckian and Albian consists of the rare specimens discovered in the Wealden of England. Two mammals found in the Husin Series of Manchuria also may be of post-Purbeckian and pre-Albian age, but the correlation of the strata in which they were found is still open to question (see Patterson 1956, pp. 30-31). The first specimen considered to be a mammalian fossil discovered in the Wealden of England was described in 1891; in the two following decades four more specimens were found. After this initial activity almost fifty years elapsed before another Wealden mammalian fossil was discovered. In the autumn of 1960, working under the auspices of a National Science Foundation Postdoctoral Fellowship, I began a survey of British Mesozoic mammals and the localities at which they were found. At the suggestion of Dr. K. A. Kermack, one of the first localities visited was Cliff End, where three of the Wealden mammalian fossils had been found in a bone bed cropping out on the shore. Blocks of the bone bed found scattered along the beach proved to contain mammalian fossils. As word of my interest in Wealden mammals spread, I was told of another promising bone bed in Paddockhurst Park that also proved to contain isolated mammalian teeth. Some of these fossils were exhibited at a meeting of the Geological Society (Clemens 1960). This paper is an interim report on a research project that is being continued under the direction of Dr. K. A. Kermack of the Department of Zoology, University College, London.
All catalogue numbers cited are from the British Museum (Natural History) Catalogue of Fossil Mammalia.
In 1891 A. Smith Woodward announced the discovery of a tooth of a Wealden mammal. Two years later Richard Lydekker described another specimen and in 1911 Woodward reported on three more. Study of these fossils and the circumstances of their discovery has brought to light facts confirming the unpublished suggestions of several palaeontologists that the identification of some of the fossils as teeth of Wealden  mammals is incorrect. The first fossil thought to be a Wealden mammalian tooth (M13134) was found by Charles Dawson in a bone bed within the Wadhurst Clay exposed in the 'Old Roar Quarry' near Hastings. It was an isolated tooth having both the heavily worn crown and root preserved. Later it was damaged and all that remains is a fragment of the root. Woodward (1891, p. 586) described the crown of the tooth as being '. . . supported by two roots, which are robust, of nearly equal size and depth, incompletely separated in the upper half and terminating obtusely'. Loss of parts of the tooth has revealed that, although deeply grooved on the surface, the root contains a single pulp cavity rather than two as would be expected if this structure were two incompletely separated roots.
After comparing this Wealden tooth (M13134) with haramiyid teeth and the M2 of Ctenacodon (Plagiaulax )minor, Woodward tentatively identified it as the molar of a multituberculate and allocated it to the genus Plagiaulax under the provisional name P. dawsoni. The similarities on which this identification was based are neither numerous, detailed, nor fully convincing. Simpson (1928, p. S2) commented that the tooth '. . . is probably Plagiaulacid, although its generic and specific afflnities are quite undeterminable'. The presence of a single pulp cavity in the root reduces the probability that the tooth is a plagiaulacid molar, which has two roots, and strongly suggests that it is not the tooth of a mammal.
In 1893 Lydekker identified another fossil (M5691) as an incisor of a Wealden multituberculate. In the discussion following the presentation of Lydekker's paper at the Geological Society, Sir John Evans recounted the history of the discovery of the tooth: 'He [Evans] found it at Hastings, in a block of Tilgate Grit which formed part of a heap by the side of the sea-shore, and almost immediately afterwards gave it to Prof. Prestwich, in whose collection it was mislaid for a period of over thirty years. On again coming across it, Prof. Prestwich placed it at the disposal of the speaker . . .' (Lydekker 1893, p.283). In the course of the same discussion, Dawson stated: ' . . . it was unfortunate that the specimen had been taken from a loose block, because at Hastings stones foreign to the district and miscellaneous drifted stones from the shore were frequently broken up for road-metal. From the limited view permitted him of the specimen that evening he was unable to identify the matrix as from the Hastings district; and he did not recognize the fragmentary specimen before them as a portion of a mammalian tooth.'
Dawson appears to have been alone in questioning the identification of this fossil as a mammalian tooth. Study of the specimen substantiates Lydekker's observation that there is no way of distinguishing it from a rodent incisor. Because it was and still is thought to be unlikely that the Rodentia had evolved as a discrete group in the Early Cretaceous and, because multituberculates possessed somewhat rodent-like incisors, Lydekker and others concluded that this tooth was probably a multituberculate incisor. Excluding the specimen in question, the available fossil record of multituberculates demonstrates that incisors closely resembling those of rodents were evolved only in one Tertiary subfamily, the Eucosmodontinae. The convergence was far from complete and isolated incisors of these and other multituberculates can be distinguished from the incisors of rodents (Jepsen 1937). On the basis of the size, curvature, and cross-section of the crown and the distribution of enamel; the tooth described by Lydekker can be identified as an incisor of a rodent. The only attribute of the fossil that precludes this identification is its supposed stratigraphic provenance. The fossil was preserved in a  block of stone picked up on the shore. Its stratigraphic provenance was determined through identification of the enclosing sediment as 'Tilgate Grit', a name once applied to calcareous sandstones now known to be present at various levels within the Hastings Beds. This identification was challenged by Dawson and later Simpson (1928, p. 192) commented: 'The specimen, moreover, is not in such a sandstone [Tilgate Grit] but in a very crumbly clay, so that its exact level must be considered very dubious.'
To the best of my knowledge no teeth closely resembling the fossil described by Lydekker have been found in situ in English Wealden strata. The tooth from Hastings is similar in morphology and, superficially at least, in mode of preservation to upper incisors of Theridomys sp. found in argillaceous, Tertiary strata cropping out on the Isle of Wight and adjacent parts of the mainland. The nearest outcrops of these strata are over 70 miles away from Hastings. Similar incisors of rodents also have been found in Tertiary strata of the London basin that extend to within approximately 35 miles of Hastings. Dawson was first to suggest that the fossiliferous block might not have come from the Wealden cropping out near Hastings but might have been introduced from some other area. Although well removed from the Hastings district, the Tertiary strata of the Hampshire and London basins are not so distant as to preclude all possibility of the specimen being accidentally introduced from either area.
In summary, if an identification were based on dental morphology alone the tooth in question would be identified as a rodent incisor. The only controverting evidence is the supposed stratigraphic provenance of the fossil. The identification of the enclosing sediment as Tilgate Grit is incorrect, and there is no unequivocal evidence to support the former assignment to the Early Cretaceous. The block containing the fossil possibly was a fragment of a Tertiary stratum drifted on to the beach at Hastings or accidentally introduced during road construction. Although final judgement should be reserved until our knowledge of the morphology of Wealden multituberculates is more complete, the available evidence suggests that the fossil (M5691) probably is the incisor of a Tertiary rodent rather than a Wealden multituberculate.
In 1911 Woodward described three more specimens that had been discovered at Cliff End by P. Teilhard de Chardin and F. Pelletier, who were assisting Dawson. Two of these fossils are multituberculate teeth. One of them (M10480), the type of Loxaulax valdensis Simpson, had been described in detail by Woodward (1911, pp. 278-9) and Simpson (1928, pp. 49-50). The second (M10481) is probably a fragment of an M2 of a member of the same species. Both teeth are still imbedded in sediment that is lithologically similar to the blocks of Cliff End Bone Bed found by our field parties. The third tooth was described by Woodward (1911, p. 278) as follows: ' . . . most of the crown has decayed, but the two divergent roots are well displayed, the one somewhat stouter than the other'. This fossil was not recorded in the British Museum (Natural History) Catalogue of Fossil Mammalia in 1912 when the other two teeth from Cliff End were assigned specimen numbers. Recently a specimen was found in the collection with the following label: 'Mammal tooth, Fairlight, Ashdown Sand, N. Hastings, C. Dawson 15/1/1911.' This fossil, now catalogued as M20241, agrees with Woodward's brief description and probably is one of the fossils discovered by Teilhard de Chardin and Pelletier. The crown is almost completely destroyed; only a small band of enamel or enamel-like substance remains. The coronal dimensions are: length = 3.1 mm., width > 1.2 mm. The root consists of a large, cylindrical, grooved body approximately 1.8 mm.  long that terminates in two divergent branches, the largest of which is approximately 1.1 mm. Iong. Even though the root is bifid epically it does not resemble the roots of Late Jurassic or Early Cretaceous mammalian teeth. Also the teeth of these mammals are, for the most part, smaller than M20241. The mammalian affinities, if any, of this tooth will have to be regarded as indeterminable.
Thus of the fossils collected in and prior to 1911 only M10480 and M10481 remain as unquestionable Wealden mammalian teeth. The mammalian affinities of two of the other fossils, M13134 and M20241, are doubtful, and the tooth described by Lydekker, M5691, probably was not part of the dentition of a Wealden multituberculate. After 1911 no more Wealden mammalian fossils were discovered until 1960; but interest in the stratigraphy and the fauna of the Wealden did not decline. Of the studies carried on after 1911, the work of P. Allen, especially his reports on the stratigraphy (Allen 1959) and a study of Wealden bone beds (Allen 1949), and Simpson's (1928) review of Wealden mammals, are especially pertinent.
DESCRIPTIONS OF THE LOCALITIES AND THE MAMMALIAN FOSSILS
Grid coordinates given in the locality descriptions are those of the National Grid and lie in the 100-kilometre square 51 (TQ).
(1) CLIFF END BONE BED
Location and collecting technique. The exact locality where Teilhard de Chardin and Pelletier collected was not described in detail by Woodward (1911). Many years later Allen, using a description of the site obtained from Woodward in 1938, relocated the bone bed and published a locality description (Allen 1960a, p. 11) in which he stated: '. . . rapid erosion of the cliffs has re-exposed the celebrated bone-bed . . .'. On our collecting trips in the autumn of 1960 and the following spring an outcrop of the bone bed could not be found; apparently the forces of erosion had destroyed or covered what they once revealed. Loose blocks of bone bed containing the new mammalian fossils were, however, collected from the wave-cut platform and the shingle; most were found within 100 yards of the foot of the cliff. These blocks ranged in size from cobbles to a fragment of the stratum nearly 4 feet in diameter and were distributed over an area (text-fig. 1) extending northward from a point approximately 2,000 feet north-east of Haddock's Cottages (88751280) to a point near the northern end of the cliff (88851305). Because no bone beds could be found cropping out in the area, it is assumed that all the loose blocks were derived from the intermittently exposed Cliff End Bone Bed described by Allen.
The Cliff End Bone Bed is a thin stratum composed of the comminuted bones and teeth of fish, amphibians, reptiles, and, very rarely, mammals concentrated in a hard, calcareous, coarse sandstone. The stratum, part of the Ashdown Beds, is never more than 4 or 5 inches thick, and its upper surface is ripple-marked. In order to find the minute mammalian fossils a large quantity of rock, 2 or 3 hundredweights, had to be processed. After some experiment the following technique was found to be most efficient: The calcareous cement was dissolved with dilute formic acid (a 10 to 20 per cent. solution). The material that remained was washed and dried and then passed through a set of graded sieves. The coarser fraction, material caught in 12-mesh and coarser sieves, was sorted without further treatment. The finer fraction was placed in a mixture of bromo