Iviii ITINERARY. 



ovei' nearly the whole area, except to the extreme northern side. 

 Before this was an enormous gorge, always filled with rising mist ; 

 and it was clearly the channel of the Kotinga stream, into which 

 many of the valleys discharged their water from far invjards beyoml 

 the middle line. Representing this in a way, but in the othin- 

 direction, was a wide, open, long vallej', sloping slightly from the 

 eastern edge towards the western, past our camp wheie it spread out, 

 and lying not very far from and roughly parallel to the south-western 

 face by which we had come up. In wet weather its waters would 

 partly help to swell the various falls over the edge of the great oliif 

 and also the stream that would discharge from the gorge at the top of 

 the ledge. It afforded the easiest walking on the plateau, and is seen 

 in part in the illustration opposite the map in Vol. 1. There was 

 entrance from it to numerous side valleys running towards the 

 northern gorge, and no doubt in future ages it will be worked down 

 to a depth comparable with that of the Kotinga to-day, which, when 

 the outer part has been entirely worn down and broken away, will 

 leave an approach to the plateau not very unlike in character, though 

 different in position from that which gives access to-day. 



The fauna is small and insignificant, both as to number of species 

 and of individuals ; but with the exception of a dragon-fly, a wasp, and 

 a beetle, they were all new to science, and some of them were t}pes 

 of new genera. A few. of the insects collected were immature or 

 indeterminable. Besides these there Avere a few others which became 

 spoilt, among them an earthworm and a woodlouse. Some medium- 

 sized black butterflies, though rather common, were very difficult to 

 catch over the rough ground, and one we managed to secure was 

 hopelessly damaged. Some dark yellow and pale yellow and white 

 ones, flying high in the wind, never gave a chance of capture. They 

 seemed to be passing over rather than to belong to the Roraima 

 top. With these exceptions the specimens were described by various 

 naturalists in the ' Transactions of the Linnean Society,' 1900, already 

 referred to for the new prawn. Those obtained on the first' trip had 

 been described in the ' Annals and Magazine of Natural History,' 1895. 



Of vertebrates there were a small rodent {Rhipidomys) which v\as 

 named after McConnell — it had been caught at night in a bucket of 

 water leaning against the rocks : a Zonotrichia, also appropriately 

 named after McConnell, as, after grea,t trouble, he had shot the 

 specimens flying about the scanty vegetation : a small lizard, caught 

 under some rocks ; and a tiny black toad (yellow on the under side) 

 which was the type of a new genus, Oreophrynella, and of which 

 several specimens were- collected, as it was the commonest of all the 

 forms, though very diificult to detect among the small- pieces of 

 blackened I'ock, as well as on the general surface. Another new species 

 of the genus (0. Macoonelli) was obtained at the foot of the mountain. 

 Another form had also been collected. No fish were to be seen in any 

 of the pools, though at first glance some of the dragon-fly ]arv« had 

 been mistaken for them. A large white hawk, perhaps a species of 

 Leucopternis, was seen flying near the top, but almost certainly it 

 belonged to the slopes, where there would be more food than on the 

 practically barren top. We had thought an armadillo might be found, 



