D. M. S. Watson — The Beaufort Beds, South Africa. 391 



remarkable features detailed above is a difficult one. Bain in his 

 first paper recognized that they were not of marine origin because 

 of the entire absence of all salt-water shells. He considered them 

 of lacustrine origin, an explanation which seems to have satisfied 

 most observers to the present day. Stow, however, seems to have 

 been doubtful about it. 



It is interesting to compare the features of the Beaufort Beds 

 recorded above with those of the Tertiary deposits of the Great 

 Plains region of North America, also formerly considered as lacustrine. 

 With suitable modification the following quotation from Leidy 

 relating to them will apply to the Beaufort Series: "Whilst the 

 geological formation makes it appear that the fossils were deposited 

 in ancient lakes, or in estuaries or streams connected with the latter, 

 it is strange that they exhibit no traces of fishes or of aquatic 

 molluscs intermingled with the multitude of relics of terrestrial 

 animals. . . . Even mammals of decidedly aquatic habitat are 

 absent ; with the exception of the shore-living rhinoceros and the 

 beaver, no amphibious mammals have been discovered. Whilst the 

 fossil bones are in perfect preservation, their original sharpness of 

 outline without the slightest trace of erosion indicates quiet water 

 with a soft muddy bottom." The work of Matthew, Fraas, and 

 Hatcher has shown conclusively that these deposits are not of lacustrine 

 origin, but were laid down on vast flat plains as flood-plain and river- 

 channel deposits modified by wind-action. Matthew has shown that 

 the river-channel sandstones contain the remains of forest- and river- 

 living animals, whilst the fine clays enclose the skeletons of plain- 

 living types. 



The whole of the lines of evidence used by these geologists applies 

 viutatis mutandis to the Beaufort Series. The vast majority of the 

 Karroo reptiles are dry-land types. Dicynodon, which occurs throughout 

 the entire series, is a large and very diverse genus, into which the 

 Scotch Gordonia could be put ; some of the South African types are, 

 in fact, very similar to that form. Gordonia occurs in a sandstone 

 all the grains of which are large and rounded, and in which the pebbles, 

 occurring in thin impersistent bands, are all of characteristically 

 wind-cut shapes ; in fact, there is every reason to suppose that 

 Gordonia is a desert animal. There can therefore be no doubt that 

 Dicynodon is also a dry-land type. 



Besides the Pelecypods and fishes, the only types that we have 

 any reason at all for regarding as aquatic are Lystrosaurus and the 

 Stegocephalia, the presence of lateral line sense-organs on the skulls 

 of which shows that they must have been to some extent water-living. 



I have pointed out above that the remains of fishes are apparently 

 confined to the stratified shales and sandstones, and that the 

 Stegocephalia, which are never common, are proportionately less rare 

 in the sandstones. 



Of the only two Lystrosaurus localities I visited, one was in very 

 obviously bedded sandstones, the other in sandstone and not so well- 

 bedded shales. The thick mass of sandstone referred to above as 

 occurring on the Great Winterberg is apparently in the zone which 

 yields this reptile. 



