Francis F. Ojany: The Mound Topography of theThika and Athi Plains ofKenya 269

u. aus anderen Teilen der Siidost-Sahara erzielten gleiche mib Siidwestafrikas. Dusseldorfer Geogr. Vortrage Er orter. S. Breslau. Ergebnisse und erharten das gefundene Schema des 68-78, Maull, O. (1958): Handbuch der Geomorphologie. 2. Aufl., klimatisch-morphologischen Raumgefuges. Wien. W. in zentralen Sa Zitierte Literatur Meckelein, (1959): Forschungen der hara. Braunschweig. Budel, J. (1954): Sinai, ?die Wiiste der Gesetzesbildung", Mortensen, H. (1927: Der Formenschatz der nordchileni als Beispiel fiir die allgemeine klimatische Wustenmor schen Wiiste. Abh. Ges. Wiss. Gottingen, Math.-Phys. phologie. Mortensen-Festschrift, S. 63-85, Bremen Kl. N. F. XII, 1. Berlin. Horn. Nachtigal, G. (1879): Sahara und Sudan. 2 Bd., Berlin. Capot-Rey, R. (1961): Borkou et Ounianga. Inst. Rech. Pachur, H. J. (1966): Untersuchungen zur morphoskopi Sahar. Mem. No. 5, Alger. schen Sandanalyse, Berliner Geogr. Abh. H. 4, Berlin. De Burthe D'Annelet (1932): A Travers l'Afrique Fran Passarge, S. (1927): Die Ausgestaltung der Trockenwusten caise. Du Cameroun a Alger. Carnets de Route. Paris. im heiften Giirtel. Dusseldorfer Geogr. Vortrage u. Er De Lestang, J. (1965): Mitteilung iiber das Erdi-Becken. orter. S. 54-66, Breslau. Vortrag Jahrestagung d. Deutsch. Geolog. Ges. Hanno Quezel, P. (1964): Carte Internationale du Tapis Vegetal, ver (im Druck). Feuille Largeau au 1 : 1 000 000. Publiee par le Gou Dubief, J. (1952): Le vent et le deplacement du sable au vernement du Tchad, I.G.N. Paris. Sahara. Trav. Inst. Rech. Sahar., T. VIII, p. 123-164, Schiffers, H. (1957): Die innersaharische Landschaft Alger. Borku und ihre Bewohner. Die Erde, Jg. 88, H. 2, - (1959, 1963): Le Climat du Sahara. Inst. Rech. Sahar. S. 81-109, Berlin. Mem. T. I, T. II, Alger. Tilho, J. (1920): The Exploration of Tibesti, Erdi, Bor Hagedorn, H. (1966): Landforms of the Tibesti Region. kou and Ennedi in 1912-1917; The Geogr. Journal, 56 In: South-Central Libya and Northern Chad. A Guide p. 81-99, 161-183, 241-267, London. to book the Geology and Prehistory ed. by J. J. Wil Wacrenier, Ph. (1958): Notice explicative de la Carte liams - - and E. Klitzsch. Petroleum Exploration Society Geologique provisoire du Borkou Ennedi Tibesti of Libya, p. 53-58, Tripoli. au 1 : 1 000 000. Direction des Mines et de la Geolo Hovermann, J. (1963): Vorlaufiger Bericht iiber eine For gie de PAEF Brazzaville. schungsreise ins Tibesti-Massiv. Die Erde, Jg. 94, H. 2, Walther, J. (1900): Das Gesetz der Wustenbildung in S. 126-135, Berlin. Gegenwart und Vorzeit. Leipzig. Kaiser, E. (1926): Die Diamantenwuste Sudwestafrikas. Nachtrag: Nach Drucklegung dieses Beitrags ist erschienen: 2 Bde. Berlin. Mainguet, M. (1968): Le Borkou. Aspects d'un model - (1927): Ober Wiistenformen, insbesondere in der Na eolien. Ann. de Geogr., 421, p. 296-322.

THE MOUND TOPOGRAPHY OF THE AND ATHI PLAINS OF : A PROBLEM OF ORIGIN

With 1 figure and 2 plates

Francis F. Ojany

Zusammenfassung: Die Kleinhugeltopographie der Thika Introduction von und Athi-Ebenen Kenya: das Problem ihres Ursprungs. Der Aufsatz untersucht den Ursprung von kleinen, aber The areas to the southeast, east and northeast of zahlreichen die auf den un Hiigeln, vulkanischen Ebenen the City of is dominated by four well-de mittelbar im Osten und Nordosten des Nairobi Flughafens fined lava plains which are known locally as the gut entwickelt sind. Fruhere zum Problem des Erklarungen Kapiti, Athi, Embakasi and Thika Plains wahrscheinlichen dieser und (named Ursprungs Kleinhiigel (Gilgais from sooutheast to von northeast). Elsewhere, the present Reste Termitenhugeln) haben sich nach Uberpriifung author has termed these the als unzulanglich erwiesen. (1966) simply plains Auf Grund der aufieren and Thika Plains. As be these und inneren Morphologie der Kapiti might expected, are the Kleinhiigel wie auch des Befundes ahnlicher vom Verfasser plains constructional features of fissure lava untersuchter Erscheinungen in anderen vulkanischen Land eruptions which filled a region which was before that schaften Nordtanzanias werden sie als umgelagerte vulka time, probably an irregular low-Lying plain with a nische erklart. Sie erinnern an die Schlammstromhiigel ge very gentle slope to the west and northwest (see vulkanischen und ihre waltigen Eruptionen Folgeerschei Gevaerts 1965). nungen, vom denen die Landschaft sich eben erst erholt. Das hier the details the areas behandelte Problem dient auch als Beispiel fiir Although of evolution of the die was more Notwendigkeit fortgesetzter interdisziplinarer Zu complex, than can be gone into in this sammenarbeit in der Forschung. paper, it is fairly safe to generalise and say that the 270 Erdkunde Band XXII

four plains mark different lava eruptions with the Previous work on theMounds Kapiti Plains (consistingmainly of phonolites), which Stephen, Bellis and Muir (1956) made a special erupted some 13 million years ago, being the oldest. of these mounds and them as The Thika Plain is almost as old, except that its study regarded being in Black Later, surface has been affected by later The gilgai phenomena Tropical Clays. eruptions. the near Tala and lavas whose surface now form the Athi Plains are Scott (1963) described examples Fairburn described those in the Estate much younger (Nairobi phonolites being only about (1963), area. On air the mounds show as island 5 million years old) while the Embakasi plains, are photographes, features in the sea of (see Plate 1). The Plate even younger still. Thus the rocks get progressively plain is a of an at the scale of younger northwestwards from the Kapiti Plains to reproduction airphotograph 1 : 25,000 that was flown in 1966, at an wards the Aberdare Ranges where the Tra January average of about 18,000 feet above the ground chytes are only about 1.7million years old. height level. It covers an area which is about two miles to From the foregoing paragraph, it is evident that the south of Kenyatta College (see actual position the four plains have been exposed to subaerial erosion marked 1 in Figure 1). Despite the great height, at for different durations so that each plain might be which the aircraft was flying, the mounds still show expected to have attained a different stage of ero clearly and the pattern which theymake can hardly sion. This in fact is the case, and the area affords a be said to fit into a arrangement. It is to particular environment in which illustrate the rela are unique nonetheless, clear from the Plate that the mounds between time, process, and rate of erosion. tionship best developed on the crests of the ridges which are The Plain, having been exposed over a longer Kapiti here by the well-developed drainage chan duration has now reached a which must be separated stage nels. described as late The Athi Plains surface is now youth. Scott described the mounds in the area which he in early youth while the Embakasi Plain is still so studied as "rising from about 1 to 2 feet above the much unaltered by the agencies of erosion, that it land surface and a diameter of about must be described as still in its initial stage. general having being observations 20 yards across". He also confirmed the The above are rimmed between the plains higher which had been made by Stephen, Bellis and Muir formed the Turoka and Kikuyu Uplands more ground by seven years earlier that the mounds support to the west and northwest on the one side, and the and better crops when under culti to east. palatable grasses older residual hills the The plains near Fairburn wrote: vation. Describing those Juja, thus form a narrow belt of topographical low which "They are usually about 20 to 50 yards in diameter, also shows as a drier corridor which receives a mean at 50 to 100 yards internal and develop a stronger annual rainfall of between 500 mm. in the south, and of vegetation than the surrounding depres mm. near Thika The growth 800 in the north, township. sions." He also noted that their distribution is not areas receive much mean adjoining higher higher "occur in crude haphazard but radiating pattern", totals 1400 mm. in the Aberdare and (over Ranges and that the mounds are usually composed of ce over mm. in the parts of Machakos to con 1,000 higher mented brecciatied material. Plate 2 helps These climatic aspects have been well-examined hills). firm some of these early observations. The photo Thompson and Sansom (1967). by graph was taken by the author by the roadside oppo site Kenyatta College (see position of Plate 2, in It to indicate the exact rela The mound topography Figure 1). clearly helps tive size and the distance apart between one mound as curve the crest On parts of the Thika and Athi Plains are devel and the next as well the gentle of a and the fact that better oped numerous little mounds which give peculiar of the mounds, they support as and puzzling relief feature known locally the vegetation. Thika and Athi Plains. The was not to con mound topography of the The present author therefore able features have engaged the attention of a number of firm that the distribution of these features conforms and the author has with pedologists and geologists present to any definite pattern. However, he concurred as a com since late 1963 examined them mainly geo the previous authors on the general size and the were found morphologist. position of the mounds. Many of them lava material of The mounds are best developed around theNairobi to be composed of agglomeratic occurrence and other all of different Falls and in the Juja area fromwhere their mixed debris fragments re extends eastwards as far as Tala Market, near Kan sizes intermingled with soil. The rock fragments To the west and sembled tuffsof theAberdare suite. It was also notice gundo in the Machakos District. are northwest of the above area, the features are also able that the sizes of these mounds much smaller east where occur on ill-drained found extensively preserved between Kahawa and further they flat, towards Koma Thika Township on either side of theNairobi-Thika parts. Around Tala Market, Rock, instead of themounds volcanic boul Road. Figure 1, illustrates this development. developing, only Francis F. Ojany: The Mound Topography of theThika and Athi Plains ofKenya 271

0 5 10 15 20 25 1 I I I I II I I I MILES Position of 1 and 2 Area wnere mounds have been Road Contours in feet Q) (2) plates mapped ^sooo"' _'_GO'S.

Fig. 1: Location and General Reference map 272 Erdkunde Band XXII

Plate 1: Air photograph showing the occurrence of the mounds ders (some loose), are found littering the surrounding gilgaies which had been reported from Australia by area inwhich much laterite has developed. Hallsworth, Robertson and Gibbons (1955). But, Scott rejected the gilgai origin for the present fea The themounds origin of tures. He noted the fact that the present mounds are too to The origin of the mounds has been the focus of big compared the so-called gilgaies reported to out even considerable uncertainty and continues puzzle fromAustralia. He furtherpointed that the many students.Thus to this end, a number of possible other gilgaies which have been reported from other more which were also Stephen explanations have been advanced. The seriously parts of Africa quoted by are advocated interpretations are re-examined below: and his colleagues, all much smaller features than can seen (i) Gilgai-origin: Stephen, Bellis and themounds which be in the Thika and Athi area. Muir, regarded these features as due to gilgai pheno River Plains Another fact which should be born to mena in tropical black clay soils similar to the other in mind here is that according the evidence from Francis F. Ojany: The Mound Topography of theThika and Athi Plains ofKenya 273

Plate 2: A close up of the mounds, near Kenyatta College

- Australia, the gilgaies there are known to be still (ii) Relic Termite Anthills" forming and that even if one is graded, itwill rebuild origin : The possibility that these mounds are more again. After than four years of observation, the "relic termite anthills" were considered by Stephen present author agreed that the Thika- and his colleagues. However, Scott discounted this are as to mounds certainly "fossil" features. possibility because he explained, it is difficult, The term "gilgai" was first used in Australia by see why termite mounds should be arranged in a Hallsworth and others (1955), and is said to be an definite pattern and also because no termites were a Aboriginal word meaning "small water-hole". found in any of thosewhich he examined in details. Other terms for similar features are "melon-hole", A number of furtherobjections can be listed against "devil-devil", and "Bay of Biscay country" (see the anthill origin. It is difficult to explain what could Prescot 1932). From the descriptions by these early have eliminated the ants. Also, in cases which the are authors, one gets the impression that the features present author examined, the internal form of these more on mounds showed no ant tunnels. The dominance of nothing than small-scale undulations the surface of the land thatmight more appropriately, be agglomerates and other large tuff boulders indicated compared in size, to pingoes that are sometimes found that the features are "in situ" and have not under in periglacial and related environments. That black gone any significant remoulding since being formed. clay soils can expand considerably under certain con This latter fact is particulary well-brought out by the ditions is not disputed here.What is questioned is the boulders which are scattered near Tala Market. Again, topographical effectsof such limited swellings. Costin if these are anthill features their general height about (1956), was probably very right when he remarked the ground would be higher and not alsways the on the similarities between gilgaies and deep-frost smooth curve which tends to be typical on most of soils. Even from the measurements of the puffs (or these mounds. Thus the internal composition and the hummocks) and the hollows (or shelves) given by smooth curve of the mounds do not support the work thosewho have examined gilgaies in other continents of ants. (see references cited), the overriding impact on scale From the foregoing consideration, the present are area Scott's conclusion that the is that they small features. In the present author favours present (see actual measurements reported by all who habe features are neither gilgaies nor relic termite anthills. studied the Kenya examples), it is thought here that What other features can they then be? It is probable the expansion of the clay minerals cannot result in that their origin is tied up with the volcanic episodes such mounds. Their diameters alone are con that have affected the region. huge - siderably greater than those of most anthills that the (iii) Volcanic mudflow mounds (1 a - present author has seen over much of East Africa, hars) origin: The present author suggests while their low height above the ground is not sug these features are probably volcanic mudflow mounds re gestive of growth from below but might be more similar to the lava mudflow which have been simply explained if one regarded them as material ported from Java by Escher(1925) and also from falling to the ground with some impact. New Zealand by Cotton (1952). Similar lava 274 Erdkunde Band XXII mounds are not unknown in other volcanic environ stributed over a long period mainly by rain expecially ments in East Africa. The present author has exam during the slightly wetter period during the Kenya ined the equally striking mound topography which "pluvial phases" of the last one million years. The is beautifully developed in the Sanya Plains, a little mudflows are here thought to have come from the way before going into Moshi from Arusha. These later eruptions of the Aberdare volcano. These erup as are to con Sanya mounds have since been interpreted lahar tions known have been accompanied by mounds associated with the Kibo lava mudflow by siderable avalanches of tuffswhich covered much of Wilcockson, Downie and others (1965). Similar theNairobi area. It was not always possible to dif mound topography was also examined by the author ferentiate the tuffs which form the bulk of the during a Department Field Class with students from mounds but many of these are thought by the present theGeography Department, University College, Nai author to be Upper Kerichwa Valley tuffs although robi in 1965 in the northeastern slopes ofMount Meru in the higher parts towards town, it was not were tra towards Ngara fromUsa. These latter type possible to say whether these Limuru or are associated with Mount Meru mudflows. To look chytes not. at, both the Sanya and Ngara Nanyuki lava mounds are very similar morphologically with the mounds Conclusion which are founds in the Thika-Athi Plains of Kenya. The morphological similarities of the mounds in mode evolution Suggested of the Thika area with those already identified from theThika and Athi Plain mounds for Northern Tanzania as being lahars mounds gives the case as It is known that the evolution of the present vol strongest for interpreting the Kenya examples lava mudflow. Also the reasons ad canic landscape around Thika was characterised by being already vanced the and "fossiltermite ant sporadic mudflow episodes. This fact, it is suggested against "gilgai" hills" seem to be If the above may give the clue for a rational interpretation of the impelling. suggestion is then it is reasonable to conclude also that origin of the Thika mounds. Some of the tuffs in the tenable, the Thika mounds are a little older than the Arusha area were probably distributed during phases of scoria Moshi eruption and deposition. It is quite usual formudflow examples. and other debris to be distributed over wide areas far Most of the previous research students who have tried to examine the of these features have from the source of the volcanic cone during such origin tended to restrict their own outlooks far too much violent phases of volcanic eruption. Cotton (1952), the research in their own pointed out that this is the origin of the Java mud by techniques disciplines. Thus have in most not contained or flow or lahars ("lahars" is a Javanese word formud they instances, into account other environmental considera flows which is now in use as a general term in vol taken tions which have a or even clues canology). might bearing give to their matter. Thus the soil scientists have Cotton also pointed out that "rain falling" in the subject course been content with detailed to the ordinary of events, expecially tropical rain, laboratory analysis exclusion of the fact that the area has may be quite sufficient to cause extensive sliding and complete only severe disturbances. re-distribution of unconsolidated ash by mudflows. recovered from volcanic Perhaps the nature of the features illustrate the Sheet wash is also known to be capable of aiding in intriguing in the forma need which is now more and more re-distributing "cold-lahars" and help only becoming in re tion of the mound or hummocky landscape (lahar recognised of inter-disciplinary co-operation areas to search.Here the of the soil the landform) in close large volcanoes. techniques pedologist, are the an indeed the It is possible therefore that the present mounds geologist, geomorphologist geochro It is that the contribution lava mudflow mounds similar to those described by nologist. hoped present some new to Cotton fromNew Zealand, and to the "ten thou may throws light the better understand of the features of our environment. sand hillocks" of Tasikimalaja in Java, described by ing Escher and also referred to by Cotton. This view is strengthened by the occurrence of lahar mounds in Acknowledgements Northern Tanzania. Besides those already referred The author wishes to record his to the to by the present author, others have also been noted gratitude Nairobi for the financial assist by the author in the Southern end of the floor of University College, have ance towards the cost of the illustrations and field Ngorongoro Crater. The Ngorongoro examples work. His thanks also to theDirector of clearly been resorted by rain and sheet-wash. The go Surveys, to use No. 750 Kibo and Meru lahars have been described by Downie Kenya for permission airphotography on as Plate he must thank and his colleagues (1965) in the Explanatory notes (here reproduced 1). Lastly, those of his friendswho have made useful comments theGeological Map of Kilimanjaro. A view is advanced here that in the Thika and and suggestions to the paper. The shortcomingsmust Athi Plains area, the mudflows have been re-di however remain his alone. Albrecht Kefiler und Felix Monheim: Wasserhaushalt des Titicacasees 215

References 8. Preston, J. A. (1932): In C.S.I.R. (Australia) Bul letin No. 52. (The present author has not been able to 1. Costin, A. B. A Note on the and (1956): Gilgaies see the full length of this article). Frost Soils. The Journal of Soil Science, Vol. 6, 9. Scott, R. M. (1963): The Soils of the Nairobi-Thika pp. 32-34. Yatta-Machakos Area. Dept of Agriculture, Kenya, 2. Cotton, C. A. (1952): Volcanoes as Landscape Forms. pp. 14-16. Whitcombe & Tombs, 2nd Edition, pp. 214-253. 10. Stephen, I., Bellis, E., and Muir, A. (1956): Gilgai 3. Escher, B. G. (1925): L'eboulement prehistorique de phenomena in Tropical Black clays of Kenya. The Tasikimalaja et le volcan Galounggoung, Java. Leid Journal of Soil Science, Vol. 7, pp. 1-9. sche Geol Meded., Vol. 1. pp. 8-21. 11. Thompson, B.W., and Sansom,H.W. (1967): Climate. 4. Fairburn, W. A. (1963): Geology of the North In Nairobi: City and Region. Edited W. T. W. Machakos-Thika Area. Report No. 59, Geol. Surv. by Oxford Press, Nairobi, pp. 20 bis Kenya, pp. 25-26. Morgan, University 38. 5. Gevaerts, E. A. L. (1964): Hydrology of the Nairobi 12. W. and others Area. Govt. Kenya, Water Development Dept. Tech Wilcockson, H., Downie, C, (1965): Notes on the of Kili nical Report No. 1, pp. 34, see Chapter 3, p. 4. Explanatory Geological Map Geol. Surv. Division. 7. 6. Hallsworth, E. G., Robertson, G. K., and Gibbons, manjaro. Tanzania, p. F. E. (1955): Studies in Pedogenesis in New South Note: A discussion of the problem of gilgaied soils in Wales, VII: The ,Gilgaic Soils. The Journal of Soil the recent literature is given in Bremer, Hanna: Muster Science, Vol. 6, pp. 1-31. boden in tropisch-suptropischen Gebieten und Frostmuster 7. Ojany, F. F. (1966): The Physique of Kenya: A Con boden. Annals of Geomorphology, N.F., Bd. 9, Berlin tribution in Landscape Analysis. Annals Assoc. Ameri 1965, pp. 222-236. can Geographers, Vol. 56, pp. 183-196. The Editors

DER WASSERHAUSHALT DES TITICACASEES NACH NEUEREN MESSERGEBNISSEN

Mit 3 Abbildungen und 4 Tabellen

Albrecht Kessler und Felix Monheim x)

water new Summary: The budget of Lake Titicaca, after In seinen Untersuchungen zur Klimatologie und measurements. Hydrologie des Titicacabeckens zeigte F. Monheim Since 1956, there have been 37 precipitation gauges in 1956, dafi in der bisherigen Literatur sehr unterschied the Titicaca Basin and from (Peru Bolivia). Apart these, licheAuffassungen iiber den Wasserhaushalt des Titi measurements have been carried out since 1957 discharge cacasees und iiber den Abflufi durch den Desagua on the most important streams flowing into Lake Titicaca dero bestehen, der z. B. von A. Forti zu 140 mVsec and on the Desaguadero. In addition, both authors have berechnet wurde, von F. Monheim aber nur zu 20 m3/ carried out their own observations. With the help of all Die these a map of annual precipitation for the 1957-1961 sec2). damaligen Wasserhaushaltsberechnungen beruhten freilich auf unvollkommenen da period was prepared and, using the water budget balance, Unterlagen, evaporation, until now only estimated, was calculated. der Zuflufi zum See noch nicht gemessen wurde und The picture of precipitation distribution is particularly auch die Niederschlage wegen derWeitmaschigkeit des with a maximum of 1150 mm over surprising, pronounced Beobachtungsnetzes und manchmal auch wegen man the northern of Lake in part Titicaca, partly explicable gelnder Schulung der Beobachter nichtmit der notigen terms of relief (lower height of the east Cordillera be Genauigkeit bekannt waren. tween Cordillera Real and the Apolobamba cordillera). Above all, however, the nightly land wind with conver Im Zusammenhang mit den schon langer disku gence of ground winds over the lake and higher humidity tierten Projekten zur Nutzung des Wassers des Titi there, contribute to this maximum. cacasees zur zur Elektrizitatsgewinnung und Bewas The water budget of the lake was calculated for the serung in der Kiistenwiiste haben nun five individual years 1957 to 1961. The moisture receipts pazifischen Peru und Bolivien seit 1956 zahlreiche neue meteoro of the lake are accounted for 58 ?/o by precipitation and Stationen im 42 ?/o by surface inflow, the losses 2 ?/o from discharge and logische Titicacabecken errichtet.Aufier 98 ?/o (1480 mm) from evaporation. The mean annual level dem werden in beiden Landern seit 1957 Ablufimes of the water budget components shows that evaporation sungen in den wichtigsten Zufliissen des Titicacasees has the lowest variation in to monthly comparison preci und im Desaguadero durchgefuhrt. Dieses Material pitation and inflow. In conclusion, the water budget of konnten die Verfasser 1962 auf einer the Titicaca Basin is examined and a mean co gemeinsamen discharge Forschungsreise einsehen und dann 1966 auf der La efficient of 0,21 calculated.

*) F. Monheim: Einleitung und Niederschlag; A. Kess ler: Zufliisse und Wasserhaushalt. 2) F. Monheim, 1956, S. 94.