TRANSACTIONS OF SECTION C. 713 



among the hills bordering the maritime plain. The valleys contain a great deal 

 of ' valley fill,' which at once absorbs the water, presiu-ves it from evaporation, 

 and good quantities can be obtained from wells. 



In the ' cotton soil ' area water is only found where hills of decomposable rock 

 break the impermeable layer, at the same time forming a porous mass which holds 

 the water. 



Where sandstone, and not crystalline rock, underlies the surface deposits, the 

 underground water is, perhaps, mainly supplied by seepage from the rivers, and is 

 not dependent on the ability of rain-water to penetrate the surface deposits. 

 Though the water-level is often more than 100 feet beiow the surface, water 

 may be obtained for some time after the rains from sliallow holes made in the 

 beds of khors, and it does not appear that an im])ermeable substratum is necessary 

 to maintain this upper zone of saturation. 



5, Contemjjoraiieous Erosion in, the Loiver Series of Goal Meanuren of the 

 Bristol Coal-field. By H. Boston, 



0, Keport on the pre^Devonian Beds of the Mendips and the Bristol 



Area. — See Reports, p. 286. 



7. On a Fossil Reptile with a Trunk from the Upper Karroo Rocks of 

 Cape Colony. By Professor H. G. Seeley, F.R.S. 



In all the reptiles from South Africa hitherto made known the external 

 openings of the nostrils are divided, with the possible exception of the genus 

 Gorgonops — a Theriodont reptile, in which the nasal openings are terminal and 

 have the aspect of the figure go compressed from above downward. The animal 

 now described belongs to the Bidental or Dicynodont type, in most of which the 

 uares are at the sides of the face, as in existing lizards, and are sometimes far 

 away from the end of the snout, as in VtijchognatJius. The animal is shown to 

 belong to this group by having two large lateral tusks or canines, like all the 

 genera of the family, and by the palate making no divergence in plan from that of 

 Dicynodoii. The palate forms a regular arch from side to side. Above it is 

 the slightly reniform nasal opening, 4j inches wide and an inch deep. The 

 aperture is unlike that in the skull of a tapir, and closely similar to that in the 

 skull of an elephant to which the trunk is attached. Its borders are smootrf on 

 the inner side. At the bottom of this excavation the cavity contracts to half the 

 width, but is not divided into two nostrils, though the vomer appears to rise, sj 

 as to indent the base of the nasal passage. The author regards the teeth oi 

 Dicynodonts as having been used, probably, to pull down structures like ant-hills 

 formed by the white ants, and finds in the narrow groove between the rami of the 

 lower jaw in Dicynodonts a channel which carried the extensible tongue adapted 

 for an insectivorous diet. In this case the large teeth were equally adapted for tear- 

 ing open aut-hills. But the remarkable evidence of a trunk indicates that the head 

 was extended in front, by a soft flexible part which would give it a functional 

 elongation, like that of the Orycteropus or Myrmecophaga, The aperture which 

 indicates the trunk is so like that in the skulls of elephants as to suggest a trunk 

 of a similar kind, and only difiers in the absence of any depth of bone between 

 the mouth and the nose such as in elephants carries teeth. The skull is 13 inches 

 long, and appears to have been 14 inches wide behind. It is not perfect, and 

 no other part of the skeleton is known. It is proposed to name the fossil 

 Kannetneijeria proboscoidcs. It was found in 1895 by Dr. Kaanemeyer, neay 

 Jiurghersdorp. 



