fractured zone. A few narrow ones have been noted on the Ngong scarp, where they traverse basic tuff nearly vertically. There are several in the Barajai Gorge composed of the same lava as the Group VI. rocks which they traverse. Where dykes occur, they seem to tend to follow lines of conjugate weakness in preference to the main lines, or it may be that they are more evident when they do so than when they follow the main lines. The effects of denudation in the fractured zone will now be briefly referred to. As already mentioned, the present climate of the lower levels of the fractured zone is semi-arid. The natural main line of drainage is through the Kedong basin, the Barajai Gorge, and the Enkobirri Gorge. The channel of the Kedong River occupies the Kedong basin, and during wet seasons the stream flows through the Barajai Gorge and reaches the northern end of the Enkobirri Gorge, before it is completely lost by absorption and evaporation. In the dry season the flow in the Kedong River, usually amounting to rather over a cubic foot per second, emanates chiefly from springs at the foot of the Kedong scarp, and generally fails to reach the north end of Najele Hill. No flood channel exists as a surface feature in the Enkobirri Gorge, except occasionally where one has been formed by the flood discharge from a cross valley for a short distance. A continuous down grade, however, exists from the watershed line, 150 feet above Lake Naivasha in the Njorowa Gorge ('The Rift Valleys and Geology of East Africa,' Gregory, 1921, p. 168), through the Njorowa Gorge, over the Suswa plains, through the Kedong basin, the Barajai and Enkobirri Gorges, and eventually to the Kura plains in the neighbourhood of Lake Magadi (vide "The East African Trough in the Neighbourhood of the Soda Lakes," Parkinson, Geogr. Fourn., July 1914, pp. 41, 42). This line of drainage carried a large discharge in Lower Pleistocene time, for its valley is floored and flanked by lake and river sediments throughout much of its course, indicating a succession of lakes chiefly fed by the discharge from the larger and deeper Lake Naivasha of that time. Some of the sediments are older than the volcanic rocks of the sixth group. The southern portion of the Kedong basin contained one of these lakes, and its sediments yield water-worn pebbles of a kind of lava (comendite), which is not known to occur in situ nearer than the hills south of Lake Naivasha, where they are traversed by the Njorowa Gorge. The Barajai Gorge is the homologue of the Njorowa Gorge. It is flanked by cliffs usually varying from 200 feet to 400 feet in height, but in the case of one scarp rising to over 700 feet. It will be observed from the map that the course of the Kedong River, flowing in a south-south-west direction along the foot of a steep scarp in Group VI. lava, is deflected to the west at the northern foot of Najele Hill and follows a westerly course for 2 miles along a plane of weakness, probably engendered by conjugate jointing and slip faulting; it then meets one of the meridional fracture lines and Enkobirri gorge looking S.W., Lacustrine sediments in foreground and middle distance Lake sediments at Kichero in the Kedong Basin, looking E. Kedong main scarp in background follows it for a mile, when it is again deflected along a short cross valley to the Enkobirri Gorge. The Barajai Gorge and much of the Enkobirri Gorge appear to have been eroded in Lower Pleistocene time, but previously the discharge from the northward flowed along different lines at higher levels; for the remains of sediments, terraces, and water-eroded rocks indicate that, prior to the formation of the Barajai Gorge, the main line of drainage followed, at different times, adjacent meridional valleys having floors several hundred feet higher than that of the existing Barajai Gorge. One of these was the valley east of Najele Hill, and another the valley east of Tiridik. Prior to the subsidence of the higher elevations, within the Rift Valley, of the volcanic rocks of the fourth group, along the lines of the Kedong main and cross faults, but subsequent to the development of the structure of the existing grid, the southern discharge from these higher elevations must have been carried down the parallel valleys of the grid, widening and deepening them. Vestiges of river terraces, although largely obliterated and rearranged by wind action in a semi-arid climate, exist; and the occurrence of water-worn pebbles of a type of lava similar to one which occurs in the Aberdare Range, which bounds the Rift Valley 40 miles to the northward, indicates the possibility of the drainage from the southern end of the western flank of that range having then been carried into the Rift Valley, instead of away from it as at present. Much of the earlier discharge of the higher levels of the eastern flank of the Rift Valley must have been conveyed past Naironde, west of Ngong Hill, through the deep valley of Ol Keju Lolosogwen, to the plains of Ladariak, west of the Kapiti scarp. Some of it, however, seems to have terminated in enclosed lakes and swamps within the Nairobi District. The intense precipitation which occurred in East Africa during Lower Pleistocene and earlier times, when lakes, now dried up, existed and existing lakes were much extended, has been referred to by numerous writers, notably, in respect of Kenya Colony, by Gregory, Hobley, and Parkinson. During this period the glaciation of Kenya Mountain extended to 4500 feet below its present level. In the Athi River Valley, within the undisturbed zone of the Nairobi District, one or more shallow lakes of considerable extent existed, the sediments of which are indicated on the map accompanying this paper. Mr. C. E. P. Brooks has shown (Quart. Fourn. Met. Soc., vol. 40, 1914) that high rainfall in the tropics during the Glacial Period was a necessary consequence of the large polar anticyclones of that period. It will be seen that the parallel valleys of the grid-structure are regarded as having been initiated, and given direction, by gravity faulting and jointing in a region of east-west tension, and as having been widened and deepened by erosion. Their lower parts are frequently found to be partly filled, or levelled off, by a wind-carried loess-like loam (Maufe, Report relating to the Geology of the East Africa Protectorate," 1908). The steepness of the upper levels of the scarps of the valleys seems often to be due to exfoliation, the products of which collect at lower levels of the scarps as screes at comparatively gentle slopes. This exfoliation is helped by the vertical joint planes. It is very difficult to form an opinion as to how much of the existing form of the valleys is due to each cause. In the case of the Enkobirri Gorge, the final act of erosion through preexisting sediments appears to have deepened the gorge by about 135 feet. As mentioned by Thiele and Wilson, when considering the relation between the tectonic and physiographic features of the Rift Valley zone south of the Zambesi ("Portuguese East Africa between the Zambesi River and the Sabi River," Geogr. Fourn., 45, 1915), "There is no doubt that many important surface features have been largely determined by earth-movements, but it is not always easy to decide what share is due to tectonic forces, and how much should be attributed to general erosion, guided and influenced by rock texture and structure." In treating of the Nyasa trough, Andrew and Bailey (Quart. Fourn. Geol. Soc., 60, p. 233) indicate the partial fashioning of the trough in its present form by forces other than earth-movement. In common with most of the volcanic rocks of Kenya Colony, other than the pre-Rift Valley series in the west of the Colony (Oswald, "Miocene Beds of the Victoria Nyanza," Quart. Fourn. Geol. Soc., 1914) and lavas in the north of the colony (Parkinson, Abst. Proc. Geol. Soc., 1915), all are characterized by richness in soda and generally by extreme poverty in lime and magnesia. In respect of silica content, typical specimens from the groups show, on analysis by Mr. Colet Birch, Government Analyst, Kenya Colony, the following percentages: I. (subgroup 1), 54.00; I. (sub-group 2), 56.70; II., 46.34; III., 57′53; IV., 66.34; V., 70.30. The chemical composition of the lowest component of Group I. may be regarded as representative of the primitive soda-rich magma of intermediate acidity which on extrusion produced extensive plateaux, immediately overlying the ancient crystalline rocks, which form the basement of Kenya Colony. In composition it is very similar to that of the plutonic plug (nepheline syenite) of Kenya Mountain. It would therefore appear that the original soda-rich magma, having a silica percentage of 54, as represented by the lowest member of Group I. lavas, became differentiated, and in subsequent eruptions in the district, from closely adjacent vents, gave rise to products having silica percentages of 70 at the acid end and 46 at the basic end of the series. These circumstances are not without their significance in the consideration of the physiography of the Nairobi District, which is the only justification for reference to them here. They appear to indicate that the original soda-rich lava of Group I., which flooded wide areas in the colony on a scale comparable with the basaltic eruptions of Iceland (Thoroddsen, Geogr. Fourn., 13, 1899), the Deccan and Oregon, were very hot and liquid magmatic eruptions, emanating from a depth, which |