66 DE. H. ~F. STASBlTsQ ON SUBFOSSIL 



here to employ a much more perfect method, by which the contents of the marsh over 

 a considerable area have been thoroughly explored, and had any perfect skeletons 

 existed they must of necessity have come to light. 



What, then, is the origin of this deposit which has yielded such valuable results in 

 spite of the fragmentary nature of the remains 1 A careful study of the topography of 

 the immediate neighbourhood has revealed the fact that the flow of lava which now 

 bounds the marsli on the north and west sides must at the time of its output have 

 completely blocked up the course of the River Mazy and formed a lake of considerable 

 area. The lava-flow now rises above the present bed of the river to a height of 

 50 metres and forms abrupt cliffs overhanging the river and showing the usual broken 

 columnar structure due to the erosive action of the river and prolonged weathering. 

 But there are two interesting facts which show clearly that this lava-stream once 

 impinged on the primitive I'OcJc on the left hank of the river. The first is that con- 

 siderable masses of lava remain in situ embedded among the primitive rock high up 

 above the present level of the stream (text-fig. 3, No. 2). The second fact is that 

 there exists on the right bank of the river, at the point where it now first encounters 

 the basaltic cliffs, a triangular section of Avhat once has been a promontory of primitive 

 rock continuous with the present cliffs on the left bank ; thus showing not only that 

 the river has been deflected from its course by the lava-flow, but that its original bed 

 has been entirely obliterated and filled in by the stream of lava. Under these circum- 

 stances one is driven to the conclusion that a lake must have been temporarily formed 

 and that the height of its surface must have been at least equal to that of the masses 

 of lava mentioned above and probably considerably higher, since detached fragments 

 of volcanic rock are found further up the hill-side. A tongue of lava at a lower level 

 has also descended on the east side of the present area of the marsh. 



Reference to the sketch-map (text-fig. 3) will show that the present course of the 

 river turns abruptly south at the eastern edge of the marsh. As already stated, the 

 stream is now far below the level of the floor of the alluvial deposits, but a broad bank 

 of water-worn pebbles which at this point overlies the tongue of lava forming the 

 eastern boundary of the marsh shows that formerly the river flowed at a much higher 

 level, and must, in fact, have been some 25 feet above the then floor of the basin. 



It is thus extremely probable that at least the lower layers of fossiliferous alluvium 

 were deposited in this lake. The presence of large numbers of aquatic animals would 

 thus be explained, and, conversely, some support is given to the presumption, arising 

 from anatom.ical evidence to be adduced later, that some at least of the gigantic 

 Lemuroids whose remains occur mingled with those of Hippopotami, Crocodiles, and 

 Chelonians, were themselves aquatic. In the case of the jEpyornis and other land- 

 animals, one may suppose that during the coui'se of the lengthened period which must 

 have elapsed ere the river wore down its exit sufficiently to drain the lake, numerous 

 carcases of these animals were brought down by the annual floods, and finally settled 



