Front Cover
 Title Page
 Table of Contents
 Notes on contributors
 Map of the Ankole district
 The making of Ankole
 The physiography of Mengo district,...
 The Karamojong and the Suk
 Uganda's old eastern province:...
 The fishes of uganda - III
 Food, protein and kwashiorkor
 Early days in Kampala
 Nyoro personal names
 The Rev. A. B. Fisher in Uganda:...
 Notice to contributors
 Back Cover

The Uganda journal
External Link ( Journal issues, 1999-2002 )
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00080855/00041
 Material Information
Title: The Uganda journal
Abbreviated Title: Uganda j.
Physical Description: v. : ill. ; 22-24 cm.
Language: English
Creator: Uganda Society
Publisher: Uganda Society etc.
Uganda Society etc.
Place of Publication: Kampala etc
Frequency: irregular[1976-<1995>]
semiannual[ former 1934-]
annual[ former <1948>-1973]
completely irregular
Subjects / Keywords: Periodicals -- Uganda   ( lcsh )
Genre: serial   ( sobekcm )
periodical   ( marcgt )
Spatial Coverage: Uganda
Dates or Sequential Designation: v. 1- 1934-
Numbering Peculiarities: Suspended May 1942-Mar. 1946; replaced by the Society's Bulletin.
Numbering Peculiarities: One issue published every few years, v. 38 (1976)-<v. 42 (1995)>
Issuing Body: Vols. 11-13 published in London.
General Note: Journal of The Uganda Society (formerly The Uganda Literary and Scientific Society).
 Record Information
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: ltuf - ABS8002
oclc - 01644239
alephbibnum - 000301510
issn - 0041-574X
lccn - 52026895
System ID: UF00080855:00041
 Related Items
Related Items: Uganda journal

Table of Contents
    Front Cover
        Front Cover 1
        Front Cover 2
    Title Page
        Page i
        Page ii
    Table of Contents
        Page iii
    Notes on contributors
        Page iv
        Page v
    Map of the Ankole district
        Page vi
    The making of Ankole
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
    The physiography of Mengo district, Buganda
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 20a
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 28a
        Page 29
    The Karamojong and the Suk
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
    Uganda's old eastern province: the transfer to East Africa protectorate in 1902
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
    The fishes of uganda - III
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
    Food, protein and kwashiorkor
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
    Early days in Kampala
        Page 91
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
    Nyoro personal names
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
    The Rev. A. B. Fisher in Uganda: a memoir
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
        Page 119
        Page 120
        Page 121
        Page 122
        Page 123
        Page 124
        Page 125
        Page 126
        Page 127
        Page 128
        Page 129
        Page 130
    Notice to contributors
        Page 131
        Page 132
    Back Cover
        Back Cover 1
        Back Cover 2
Full Text
U a----,,--

Vt ,, 22 o
M A C 00 5 r

,re" Mai n -o koeF O RN
y f Meg 'D w cB g rj
Th1aai n a d h kf~

0!,P 7

His Excellency Sir Andrew Cohen, K.c.m.G., Y.X.VO., 0,11.E.
President: Vicle-pmsident:
Mr. D. K. Marphatia Mr. Mark Bayrin to-n Ward
Thelresident Mr, S. W, Knlubya,'C.BE,
The Mr. J. C., D.Iawrance
The flon. Secretary Mr. M. N. Nl4ini
The Flon. Treasurer Mr. R. J. Me ta
The Hon. Editor Mr. C. N. Mukuye
The Hon. Librarian Mr. LP. Saldapha
Mi. W.Elkan Mr. K. Sempa
Mr. W. J, A. Harris Mr. P, L. Shirmie
4r Mr, J, S. Kasirye Dr. K C. Trowell, oi3.Z
Mr. E, W. Kigundu Dr. K. P. Wacbsmatin
YOM Secretary., Mr. H, F. Moths
Hon, Treasurer: Mrs. M. M. Wallis
Hon. Editors.- Mr, J, W. Pallister
Dr. A. W. Southall
fl6n., Librarian: Mr. N. D. Oram
Ron, Auditor Hon, LegalAdviser
'Mr. A.'K Stump, AAC.C.A. Mx. C. L. Holcom
carresporidingseeretary at Jinja: Mr. T. R. F. Cox, c.m.o.
Corresponding' Secretary'at Mbale: Mr. F. Lukyn Williams
Corr6,ponding Secretary at Afa$aka: Mr. E. C. Lanning
CorrespondingSecrelary at Serere; Mr. D. J. Parsons
Hon. Vice-Presidents:
Rji mutesa 11, Kabaka of Buganda Sir John Milner Gray'
R. A Sir Tito Winyi (3afabusa IV, Mr. E Haddon
CXX. Omukama of Bunyoro Mr, H. B. Thornas wi.E.
Sir E F Twining, G.c.m.G., NLB.E. Dr. A. W, Williams
Past Presidents:
03- 54 ''Sir A. R. Cook', C.M.G., O.B.f., 1,944-45 Dr. K.'ADavies, C.M.G.,O.R.E.
M`LI', 1934-35 Mr. E. J. Wayland, C.B.E. 1945- 46 Mr G. K, E, Hopkins,., o,&F-
035-46 Dr, H. H Hunter, C.BX-., LLD. 1946-47 Mrs. K. M, Trowell, m.Bx.
193E-37 Mr. H. Joyitt"C.M.G. 1941-48 Dr. W, J. Eggelin g
8, Sit H. R. Hone, 1948 Dr. G. ap Griffith
1950-51 Rev. Fr. F. B. G4ftey,'WX,
938-39 Mr. 1, Sykes, ox.R. I I 1,
9 3 9 -4 0 Mr. N. V, Brasnett 1954-52 Dr. A. W. Williams
040-41 CAptain C R, S. Pitman, 1952-53 Sir 1. B. Hutchinson,

194-1-42 Mr., S. W. Kulubya, c.B.E. 1953-54 Mr. J. D. JPMMD, OXX,
194Z-43 Mr. E. A- Temple Perkins, 1954-55 Dr. AudreyI. Richards, cR.E.
`A943-44 Mr. R, A. Snoxall t955-56 Dr, H; C. Tr6woll, oBX.
M s., K. M. Trowell, mB.E. Mr, 13. K. Mutyantil Mr. G. P. Saben
Secretary: Mrs, M, M. Wallis


Uganda Journal



No. 1

MARCH 1957


MR. J. W. PALLISTER Hon. Editors
(assisted in England by H. B. THOMAS, O.B.E.)

Published by



The Making of Ankole -
The Physiography of Mengo District, Buganda
The Karamojong and the Suk -
Uganda's Old Eastern Province: The Tran
Protectorate in 1902
The Fishes of Uganda-Ill -
Food, Protein and Kwashiorkor -
Early Days in Kampala -
Nyoro Personal Names -
The Rev. A. B. Fisher in Uganda: A Memoir


The Lango Wars with Egyptian Troops, 1877-8 -
Kibuka E. B. HADDON
The Kivu Mission, 1909-10

Sir Mark Wilson -

The Inheritance of Land in Buganda -

r to East Africa





Jinja Transformed (by C. and R. Sofer) A. LARIMORE
Bwamba (by E. H. Winter) J. MIDDLETON
The King's African Rifles (by Lieut.-Col. H. Moyse-Bartlett)

An Introduction to the Ateso Language (by J. H. Hilders and
J. C. D. Lawrance) N. C. WIGGINS




H. F. Morris. See Uganda J., Vol. 20, No. 2.

J. W. Pallister. Formerly lecturer in Geology at Birmingham University. Joined the
Geological Survey of Uganda in 1954, and early in 1957 transferred to British
Somaliland as Chief Geologist.

A. J. Docherty. Came from Palestine to Uganda in 1946. From 1947 to 1954 he was,
at various times, stationed in Moroto, Karamoja. He is now Superintendent in charge
of the central Police Station of Kampala.

W. E. Hoyle. Arrived in Uganda in 1903; was first manager of the Uganda Company
Stores on Nakasero from 1904 to 1910. He organized and became first manager of
the Uganda Bookshop from 1920 until his retirement in 1929. He is now living in

K. Ingham. See Uganda J., Vol. 20, No. 1.

P. H. Greenwood. See Uganda J., Vol. 20, No. 2.

H. C. Trowell. Is a Senior Specialist Physician in the Uganda Medical Service; he has
published many articles and books on malnutrition.

J. H. M. Beattie. Was for some years in the Tanganyika Administration. After
returning to Oxford for study he carried out field work in Bunyoro. He is now a
lecturer in Social Anthropology at Oxford.

with boundaries as established
by the Agreement of 1901.

District Boundary shown thus ... .. _-( "'"
Saza . .



S0(. "

..GO ...... ..... .... ..

/ k v.. i ... ".

Flo. 1

HE Kingdom of Ankole, as it exists today, is the creation of the British
Administration. The arrival of the British, at the end of the 19th century,
could hardly have come at a more propitious time for the Mugabe of Nkore. The
decline of Bunyoro and the break up of Mpororo had enabled a series of war-
like Bagabe, in the 18th and 19th centuries, to extend, at their neighbours'
expense, a kingdom which had previously comprised little more than the present
saza of Isingiro. On the other hand, the Baganda, also expanding, had, by the
end of the 19th century, only just begun to make their impact on the Banyankore
felt. The early administrators, therefore, found that Nkore was the most power-
ful state west of Buddu between the Katonga and the Kagera.
The first Government station was opened at Mbarara in 1898 and during the
next decade there was steady expansion westward. Supported and encouraged
by the able and not over-scrupulous Nganzi, Nuwa Mbaguta, the Administra-
tion, by force or persuasion, secured from the chiefs of the outlying areas the
acknowledgment of the Mugabe as their ruler. The name Ankole which was
given to the administrative district is a European corruption of Nkore. In fact, how-
ever, by 1901 the district was nearly twice the size of the Mugabe's traditional
kingdom of Nkore, as it had existed in 1898, and, to this day, Nkore normally
means to a Munyankore only the Sazas of Nyabushozi, Kashari and Isingiro.
The purpose of this article is to show first how Nkore grew to be the dominant
kingdom in this area and then how, under the British Administration, the
Mugabe's rule was effectively established over neighboring rulers who had
formerly been completely or partially independent. To do this I have briefly
traced the history, so far as it can be ascertained from oral tradition, of each of
the component parts of Ankole District down to the time of the Mugabe Ntare's
death in 1895, and then, in a final section, have shown, chiefly from Secretariat
records, how the early Collectors extended the District of Ankole until it
reached its present boundaries.
Nkore shares with its neighbours the legends of the famous Bacwezil kings,
though inevitably they are given a local twist. Bunyoro and Nkore traditions
agree that Nyamate was the mother of Isimbwa, the first Mucwezi, by Isaza the
last of the Batembuzi kings of Kitara, but the Nkore version adds that she was
the daughter of Ruyonga the last of an earlier dynasty which had ruled Nkore
since its establishment by the divine Ruhanga.(1) Discussion as to who the
mysterious Bacwezi were, whether Hamites as has been generally believed, or
Lwoo as Father Crazzolara maintains,(2) or Portuguese as Father Nicolet asserts
1 Throughout this article I have adopted the spelling laid down at the Runyankore-
Rukiga Orthography Conference of November 1954. As a result, many proper names
may appear unfamiliar and it should, in particular, be noted that the English ch will
appear as c, e.g. Bacwezi not Bachwezi and that the French j will, in fact, appear as j and
not as z or zh, e.g. Buhweju not Buhwezu or Buhwezhu.

with confidence,(3) is beyond the scope of this article. All legendary accounts
agree, however, that, under Ndahura the son of Isimbwa, a widespread kingdom
was established comprising Uganda and north-west Tanganyika. During the
reign of his son, Wamara, the Bacwezi were faced, so it is said, with a series of
unpropitious omens and with disobedience from their people and, in disgust,
withdrew from the land, the northern part of their kingdom of Kitara being
inherited by the Babito dynasty.
When the Bacwezi withdrew, Ruhinda, the son of Wamara and of a slave girl
Njunaki, remained behind and, having procured his father's drum Bagyendanwa,
made himself ruler of that part of Wamara's kingdom which lay south of the
river Rwizi. Although Ruhinda is considered by the Banyankore to be their
first Mugabe and the founder of the present dynasty, Nkore, or Kaaro-Karungi
(the beautiful land) as it was then called, was only a small portion of his territory
and the present ruling dynasties of Urundi, Karagwe and of many other areas in
north-west Tanganyika also claim Ruhinda as their founder. On his death, this
large kingdom split up and one of his many sons, Nkuba-ya-Rurama, inherited
Kaaro-Karungi and the drum Bagyendanwa.
Kaaro-Karungi, under Nkuba-ya-Rurama and his nine successors, was a small
area corresponding roughly to the present county of Isingiro, bounded on the
north by the river Rwizi, beyond which the Banyoro ruled, and on the south by
the river Kagera, beyond which lay the allied Bahinda kingdom of Karagwe.
In the heart of the kingdom lay Lake Mazinga (Nakivali) and it was to the
sacred forest of Ishanje on its shores that the body of each Mugabe was brought
to decompose and then to be reborn a lion. The Bagabe usually had their kraals
west and south-west of the lake in the areas of Nsongezi and Kagarama.
Throughout this period, the country was on the defensive, being subject to
periodic attacks from its powerful neighbour. The excellence of the grazing
lands and the cattle of Kaaro-Karungi was famous and the Babito Bakama of
Bunyoro who claimed, as the heirs of the Bacwezi, to be overlords of the Bagabe,
were always ready to seize an opportunity for plunder. According to tradition,
the worst invasion took place in the reign of Nyabugaro, the fourth Mugabe.
The Mukama Olimi I, having defeated and slain the Kabaka of Buganda,
Nakibinge, invaded Kaaro-Karungi, drove the Mugabe into exile, and plundered
the country. Frightened by an eclipse, so it is said, the Mukama withdrew; but
the country had been devastated. For several years cattle were so few in number
that the Bahima were reduced to living on the fruits of the earth, and men
remember these years as Eijuga Nyonza "the time of the nyonza berries for
bride price".(l)
Kaaro-Karungi entered a new phase of her history during the reign of the
twelfth Mugabe, Ntare, who came to the throne about the year 1700.2 During
2 The fixing of dates can at best be approximate. Six generations separate Ntare Kiita-
banyoro from Ntare ya Kibogo who reigned from c. 1870-1895. Allowing four generations
to a century would place Ntare Kiitabanyoro in the early part of the 18th century. Accord-
ing to Nkore and Bunyoro traditions, Ntare was a contemporary of Cwamali of Bunyoro.
Cwamali's grandfather Winyi II is said to have been a contemporary of the Kabaka
Katerega, Kabaka Juko's father, whilst Cwamali's son, Kyebambe I, was a contemporary
of the Kabaka Kagulu, Juko's grandson. This strongly suggests that Cwamali and Ntare
were one generation after Juko. If, as is generally accepted, the eclipse of 1680 occurred
during Juko's reign, Ntare's accession must have taken place about 1700.

the two centuries which followed, she became an aggressive power, expanding
northwards and westwards, mainly at the expense of Bunyoro and Mpororo.
As a young man, Ntare visited in secret the kraal of Kamurari, the Mukama of
Mpororo, and the Mukama, when he had discovered who he was, gave him two
of his daughters, Kabibi and Mukabandi, as his wives. Previously, the Bagabe
had always married into the Batwa and Baitira clans, and when Ntare returned
with his Bashambo wives, he met with considerable opposition. He was told
that rain had fallen incessantly since he had gone to Mpororo and that the royal
drums had been sounding on their own because of his marriage into a strange
clan. As a result of this and of a later dynastic alliance made by his son, the
greater part of Mpororo was to fall under the control of the Bagabe during the
18th century.
Early in his reign Ntare was attacked by Cwamali, Mukama of Bunyoro.
Utterly defeated, Ntare fled for refuge first to the caves of Nyamitsindo, then to
Kantsyore island in the Kagera and finally to Muzira forest near Nsongezi. Even
Bagyendanwa fell into the enemy's hands.3
Cwamali occupied the country for several years and then moved on to conquer
Ruanda. Here, however, the Banyoro overreached themselves. Cwamali was
defeated and killed and his army, retreating through Kaaro-Karungi, was set
upon by Ntare and decimated, Ntare thus earning the name of Kiitabanyoro.
Cwamali's mother when she heard the news is said to have cried, "Ebi shi ente
za Kaaro zankora omunda" (the cows of Kaaro have broken my heart). Thus
did Kaaro get the name of Nkore.(l)
Under Kahaya I, Ntare's grandson, Nkore's boundaries were widely extended.
As heir of his maternal grandfather, Kahaya of Mpororo, he gained control over
most of Rwampara and of Shema, whilst, north of the Rwizi, Kashari and
Nyabushozi were occupied by the Banyankore. If tradition is correct in stating
that Kahaya was the first Mugabe to undergo the ritual bathe in the pool of
Kijongo, a few miles from the present trading centre of Ibanda, then his rule
must have been effective almost to the Katonga. The Banyoro never recovered
their influence in these parts and the 18th century was to see the steady decline
of that kingdom until, by its close, the loss of Buddu and the defection of Koki
had left nothing of Bunyoro rule south of the Katonga.
Kahaya's son Rwebishengye and his great-grandson Mutambuka carried on
the work of expansion. Rwebishengye invaded and plundered Buhweju in the
west and took Kabula from the Banyoro in the east. On his northern boundary
he appointed the famous witch Murogo to act as spy on the Banyoro north of
the Katonga. Murogo and her female descendants were reputed to be able to
turn themselves into cattle and then to mix with the herds of the enemy and to
hear their secrets. They served the Bagabe for several generations in the Ibanda
area and the last of them, Julia Kibubura, was to become the only female
gombolola chief to be appointed by the British Administration.(5) Rwebishen-
gye's grandson, Mutambuka, waged wars of aggression as far afield as Buson-
gora. It was during his Busongora campaign that he captured and then married
3 On capturing Bagyendanwa, Cwamali ordered it to be cut open but, so it is said, the
perpetrator of this sacrilege was struck down by lightning whilst blood flowed forth from
the drum. Cwamali was so alarmed that he sent the drum back to Ntare (4).

the Mubito princess Kiboga who was later, when her son Ntare had come to the
throne, to have a deciding influence on the policy of Nkore.
Mutambuka died in about the year 1870 and there was the customary scramble
for power among the princes. The eldest son Mukwenda was proclaimed, but
there was a large body of opinion which favoured his younger brother Ntare.
Supported by his indomitable mother Kiboga, Ntare fought for several years
against his brother. Unable to defeat him, Mukwenda appealed to Kabaka
Mutesa who twice sent an army to assist him. Though defeated, Ntare eluded
his pursuers and, after the withdrawal of the Baganda, he again collected
together his supporters and, at Mugoye in Mitoma, he utterly defeated his
brother who was killed in the battle.
Ntare, like his father, is remembered as one of the great warrior Bagabe.
During his reign the Mugabe's power reached its furthest extent. The neighbour-
ing kingdoms of Buhweju, Igara and Buzimba recognized Ntare as their
suzerain, whilst further afield the rulers of Busongora, Kitagwenda and Bwera
(the present county of Mawogola) would send him presents to avert invasion.
The boundary with Buganda is said to have lain as far to the east as the
Kyogya river.
Ntare had, however, two new factors to contend with. The first was the grow-
ing power of Buganda which, since the conquest of Buddu at the end of the
preceding century, had been Nkore's eastern neighbour. During the latter years
of Mutesa's reign, two raiding parties were sent into Nkore and returned with
plunder. In 1888, the Baganda Christians, fleeing from persecution, asked Ntare
for help. Kiboga, remembering all that her son had suffered at the hands of the
Baganda, advised him to kill them, but Ntare allowed them to settle in Kabula.
Ungrateful guests, the Baganda were later to take that country from Nkore.
Ntare also allowed Baganda Muslims to settle in Bukanga with the result that
that country narrowly escaped incorporation in Buganda.
The second factor, which was to have more far reaching consequences, was
the arrival of the first Europeans in Nkore. In 1889 Stanley, returning from the
relief of Emin Pasha, passed through Buzimba, Buhweju and Nkore. Messages
of friendship were exchanged between Stanley and Ntare who could not but
feel well disposed towards one who had fought against his hereditary enemies
the Banyoro, and at Byaruha, some fifteen miles south of Mbarara, Stanley and
Busunku (called Uchunku by Stanley), Ntare's representative and first cousin,
made blood-brotherhood. Explorers such as Stanley were soon followed by
administrators. Nkore lay in a strategic position, controlling the main route
between Kabarega's kingdom and the German sphere of influence south of the
Kagera. Anxious to prevent the supply of arms to Kabarega or, worse still, the
establishment of German influence over Nkore, Lugard, in 1891, entered Ntare's
kingdom from Buddu. At Nyabushozi he met the Mugabe's envoys; the cere-
mony of blood-brotherhood was again carried out;(6) and a treaty was made
whereby Ntare accepted the protection of the Imperial British East Africa
Company and undertook not to allow the passage of arms through his kingdom.
A few years later, Ntare, alarmed by a German force which had passed through

his country,4 appealed for the protection promised. In August 1894, Major G. G.
Cunningham, therefore, attempted to see Ntare and get him to sign a new
treaty. Ntare, however, apparently had second thoughts on the desirability of
calling in the British, and procrastinated. Eventually, the Nganzi Mbaguta came
to Cunningham and signed the treaty on Ntare's behalf, Cunningham signed on
behalf of the Queen (29 August 1894).5 In a letter to the Commissioner, he writes,
"The Katikiro Magota arrived with full powers (he said). He signed the treaty
for the King, requested that a black man might be sent in future as your repre-
sentative as Ntali could not see a European. He said the king did not wish for a
post in Ankoli at present." The treaty provided that there should be peace
between the British and the Banyankore; that British subjects should have free
access to all parts of Ntare's kingdom and should have the right to possess
property and to trade there; and that Ntare would not cede any territory to, or
enter into any treaty with, any foreign power without consent.(7)
The last years of Ntare's reign were unhappy ones. Rinderpest had decimated
the cattle and smallpox was rampant, Ntare losing his son Kabumbire from the
disease. Finally, the Banyaruanda, provoked by the successful Banyankore
raids of a few years before, invaded the country. Before the Banyaruanda had
been driven out of Nkore; Ntare had died of pneumonia in the year 1895. With
the death of Ntare ya Kiboga, the period of Nkore's history which started with
the reign of his namesake, Ntare Kiitabanyoro, came to an end.

Since the history of this kingdom has already been the subject of an article
in the Uganda Journal,6 I propose to give here no more than a brief summary
for the sake of completeness.
The kingdom of Mpororo suddenly made its appearance towards the end of
the 17th century and as suddenly disintegrated. Traditional accounts allow for

4 The leader of this force, Major Wilhelm Langheld, in command of the German
station at Bukoba, records some details of this encounter in his book, Zwanzig Jahre in
deutschen Kolonien, 1909. The most cordial relations obtained between him and the
British authorities in Uganda. At Major Macdonald's invitation he spent a few pleasant
days in Kampala in July 1893; and it is safe to assume that problems on the common
boundary of their territories were amicably agreed upon. Immediately on his return to
Bukoba he set off on a reconnaissance of the western confines of the German sphere and
by the end of August he had reached Kaihura's in Bunyaruguru. On 7 September he
crossed the arm of Lake Edward to visit Fort George which he found abandoned (quite
recently) and burnt. From Kaihura's, having had Macdonald's permission, he decided to
return direct across Ankole to the Nsongezi ferry over the Kagera. On reaching Ntare's
he was met by Schingoma (whom he refers to as Ntare's Katikiro), but Ntare's warriors
got out of hand and, being attacked, Langheld moved on at once. On 4 October 1893
Schingoma came to Langheld at Kitangule, in German territory, with a tusk of ivory and
apologies from Ntare, who claimed that he was not responsible for the attack.
There is later evidence of full accord between Langheld and the Uganda authorities.
Early in 1895, in reprisal for an earlier raid, Kigeli of Ruanda had invaded Ankole.
Langheld (20 March 1895) prints a letter from F. J. Jackson (then Acting Commissioner)
"I am glad you do not intend to punish Ruanda for the raid on Ankole, which was a
counter-raid. I have forbidden the Ankole people to invade German territory." Jackson
speaks in high terms of Langheld (Early Days, p. 282-3). [ED.]
s The text is in Command Paper, Africa, No. 7 (1895), p. 131; also in Uganda J. Vol. 10
(1946), p. 128.
6 H. F. Morris. The Kingdom of Mpororo. Uganda J., 19 (1955), 204-7.

two Bakama only, Kamurari and his son Kahaya Rutindangyezi. Under Kahaya
Mpororo extended its frontiers to include all Kigezi (except the modern Bufum-
bira saza), the Ankole sazas of Kajara, Igara, Shema and Rwampara (except the
low ground south of the Rwizi) and the northern portion of Ruanda. During
Kahaya's reign close relations were established with the neighboring kingdom
of Nkore as the result first of the marriage between the Mugabe Ntare and
Kamurari's daughters and then of the marriage between the Mugabe Macwa
and his first cousin Nkazi, Kahaya's daughter. Kahaya's seven sons were unduti-
ful and he disinherited them and made his grandson by his daughter Nkazi, the
Mugabe Kahaya I whom he had brought up, his heir. Though some portions of
Mpororo continued to maintain an independent or semi-independent position
under the descendants of Kahaya's sons who established themselves as clan
heads in various parts of the Mpororo, the kingdom in general, from then on,
accepted either the Mugabe of Nkore or the Mwami of Ruanda as overlord.
Igara was the only portion of Mpororo which under Kahaya's sons had the
status of a kingdom with a drum. Though in the Mugabe's sphere of influence, it
managed to maintain its independence under the Benemafundo until the middle
of the 19th century when the Banyankore, after interference in a disputed
succession, managed to establish their suzerainty over the Bakama of Igara.
Shema came very early under the Mugabe's rule, having been given to the
Mugabe as a wedding present either by Kamurari or Kahaya Rutindangyezi.
Rwampara which was under the control of the descendants of Kahaya's son
Rukari was also subordinate to the Bagabe. Kajara, under the Benekihondwa
for the most part, was, on the other hand, within the orbit of Ruanda. The
present district of Kigezi was divided between Kahaya's descendants, the
Benerugambagye in Rukiga and Ndorwa and the Benekirenzi in Rujumbura
and Kinkizi. These areas were, from time to time, subject to raids from Ruanda
but, it would seem, never had any dealings with Nkore. At the beginning of the
present century, however, the Mugabe, as heir of Mpororo, attempted to assert
his authority over Rujumbura but, as will be seen later, without success.(8)
Buhweju, it is said, was part of the domain of Nyinamwiru the mother of the
Mucwezi Ndahura and the daughter of Bukuku gatekeeper of Isaza whom he
succeeded as king.(9) Bukuku, fearing the fulfilment of a prophecy that he
should be killed by his grandchild, had his daughter mutilated, depriving her
of one breast and one eye, and kept her in confinement. Nevertheless the
Mucwezi Isimbwa found his way to her and by him she bore Ndahura who,
like Perseus, eventually killed his grandfather and inherited his kingdom.(1) The
ruler of Buhweju, under Nyinamwiru's suzerainty, was Muramira a man of
great knowledge who constructed a tunnel which came out at the foot of the
Buhweju hills and by means of which he could escape from his enemies.(5)
At this time there were in Mpororo, in the present Ankole saza of Kajara,
three poor Bahima brothers of the Bariisa clan. A mystic eagle which flashed
like lightning built its nest in the brothers' house and, on consulting a sorcerer,
they were told that, if they followed this eagle, they would make their fortune.
The eagle, when it had hatched and fledged its young, flew northwards and the

brothers Kataizi, Rugo and Kinyonyi, accompanied by their sister Iremera,
followed it. Each night it would alight on a tree and the four would sleep
beneath its branches. When they reached Buhweju, Kataizi, worn out by the
journey, would go no further. His descendants became cultivators or Bairu and
the Bataizi clan remains today in Buhweju. The others continued their journey
and finally the eagle led them to the court of Ndahura in whose service they
enlisted. Ndahura, having fallen in love with Iremera, married her, and to his
brothers-in-law, Kinyonyi and Rugo, he presented the kingdoms of Buhweju
and Buzimba.(10) First, however, Muramira had to be disposed of. In this
Kinyonyi was aided by one of Muramira's chiefs who, having a grudge against
his master, betrayed the secret of the tunnel, and as Muramira emerged from it
he was set upon and killed by Kinyonyi's men who lay in wait for him.(5) In
this way did Kinyonyi, with the drum Mashaija, which Ndahura had given him,
become Mukama of Buhweju and establish the Bariisa dynasty which ruled
until the beginning of the present century.
The traditional genealogy of the Bariisa Bakama, from Kinyonyi to Ndagara
who was killed in 1901, allows for only fifteen Bakama and fourteen generations.
Yet over the same period of time, i.e. from the end of the Bacwezi dynasty, the
Baganda allow for twenty-nine Bakabaka the Banyankore for twenty-four
Bagabe and the Banyoro for twenty-three Bakama. The Bariisa genealogy of
Buzimba has, however, been even further telescoped and records only seven
generations from Rugo to Nduru, the last Mukama. Kinyonyi's successors,
though subordinate to the Bakama of Bunyoro, the heirs of the Bacwezi, and
later to the Bagabe of Nkore, had all the trappings of royalty which one expects
in the petty kingdoms of the lacustrine area. Their regalia consists, inter alia,
of the main drum Mashaija, attendant drums and nkondo7 with colobus monkey
Of the immediate successors of Kinyonyi little is known except the manner
of their death. When the Mukama of Bunyoro died the Mukama of Buhweju
would be sent for and, decked in ceremonial beads, would be led off to Bunyoro
for execution by beheading.
In the reign of Kabundami II, probably at the turn of the 17th century, the
tyranny of Bunyoro was broken. Kabundami, determined that he would not
suffer the fate of his fathers, gathered together a band of warriors known as
the Nkondami. The leader of this band was Muguta ya Butaho Enshungyera
nigacweka owa Kinika kya Katago Rubabirira Omunyanshamba ou Engabo ye
ekira ezindi,8 for such was the complete praise name of this hero, a prince of the
royal house. When the messengers came from the Mukama of Bunyoro to lead
Kabundami off to execution, Muguta took his place and, dressed in his master's
execution robes, was escorted together with his followers to Bunyoro. Here
they were bound and an axe placed before each of them. But Muguta's page
began to recite his praises and Muguta, leaping into the air, was carried into the
clouds thereupon the sky roared and the Nkondami, overpowering their captors,
beheaded them with their own axes. Then they caught and bound the Mukama
7 Ornamental head-dress.
8 Muguta the son of Butaho who leaps in the air when spears are breaking of Kinika
kya Katago the speedy one who inflicts deadly injuries whose shield excels all others.

of Bunyoro himself who, crying for mercy, promised that thenceforth the axe
would be used to cut down trees and not men.(5)
The defeat of the Banyoro by Ntare of Nkore gave Muguta his opportunity
to occupy Kashari and to conquer Nshara. Stories of the heroism of Muguta
and his Nkondami are still recited. As in all stories of Bahima warriors, cattle
is the favourite theme and the best known story is probably that of the theft of
the Mukama of Buhweju's cow, Mayenje ga Ishinjo Rutanyobwa Bigomba.
This wonderful cow which gave daily six pots of milk was coveted by Ntare
Kiitabanyoro of Nkore. The latter, therefore, sent two famous thieves Rucu
Rwa Bugoro and Runkunku to bring him the cow. By various deceits, the two
thieves managed to get hold of the cow and, having disguised its colour by
means of chalk and soot, brought it to their master. This led to war between
the Nkondami and the Mugabe's warriors the Nyana. Eventually the Buhweju
heroes were victorious, and Mayenje ga Ishinjo was returned to its owner.(5)
Stories tell of the exploits of Muguta far afield. He and his Nkondami are said
to have penetrated to Bukedi and it is near the site of Jinja that he is said to have
been fatally injured. Falling from a tree, a stake pierced and disembowelled
him. Concealing his injuries from his followers, he returned home holding in
place his rotting entrails. Entering in before his king, he expired at his feet.(l)
After the death of Muguta the lands he had won east of the Buhweju escarp-
ment fell to the Bagabe who, extending their domain over Kashari, Nyabushozi
and Mitoma, succeeded Bunyoro as the suzerains of Buhweju. There is little to
be told of the reigns of the following Bakama, Kashoma III, known as Gatuk-
wire (the red-eyed), who had been one Muguta's Nkondami, Kabundami III
and Karamagi II.
Karamagi II left two sons Rukumba and Rusharabaga. The latter was
intended by his father to be his heir but Rukumba, relying on Nkore support,
for he and the Mugabe Kahaya I had married sisters, seized the throne. Then
they took their case to Kahaya to judge but Rukumba, not content with this,
tried to spear Rusharabaga who had to flee to the drum Bagyendanwa for
refuge before escaping to Ruanda. Later, returning to Buhweju, he attacked his
brother's palace at Rukiri and, having defeated him, became Mukama in his
place.(9) During his reign Buhweju was invaded by the Mugabe Rwebishengye,
Kahaya's son.
Rusharabaga was succeeded by his son Kashoma IV and he by his son
Ndagara. Born about 1820, Ndagara had an eventful life. Early in his reign he
raided Toro and took much plunder. Then Mutambuka, Rwebishengye's grand-
son, invaded Buhweju and devastated the country. For four years Ndagara
fought with the Banyankore until, finally, he fled, first to Mpororo and then to
Ruanda. On the death of Mutambuka he returned to Buhweju.(9) During the
civil war which followed Mutambuka's death, Ntare fled to Buhweju suffering
from poison which his page had given him at the instigation of his brother
Mukwenda. He asked Ndagara for help and the latter consulted the oracles to
see whether the future lay with Mukwenda or Ntare. The answer being favour-
able to Ntare, an antidote had to be found for the poison from which Ntare was
suffering. This was to be a mixture with beer, but the country had been so
ravaged by Mutambuka that messengers had to go to Kitagwenda to get the

beer. Ntare recovered but Mukwenda sent his army in pursuit and he had to
flee from Buhweju.(4) Ndagara, however, continued to support him and when
the civil war was finally decided at Mugoye, Nyakiga, Ndagara's son, was
fighting in Ntare's army. Ndagara remained the friend of Ntare throughout the
Mugabe's reign and the Banyabuhweju accompanied him on his campaigns
against Toro, Igara, Rujumbura and Ruanda.

The story of how the two brothers, Rugo and Kinyonyi, followed the mystic
eagle to the court of Ndahura has been told in the last section. To Rugo,
Ndahura gave the kingdom of Buzimba. Some Banyabuhweju, however, main-
tain that Buzimba was at first included in Kinyonyi's kingdom, but that, when
Kinyonyi found that the area was too large for him to rule effectively, Ndahura
sent Rugo to help him. The Banyabuzimba, on the other hand, deny that
Buzimba was ever subordinate to Buhweju.
Practically no traditional accounts survive of the subsequent history of
Buzimba. The names of only seven Bakama are remembered; Rugo, Kantu,
Mukindo, Nyaruyonga, Mugarura, Nyakairu and Nduru. These Bakama, like
those of Buhweju, possessed regalia and were circumscribed by ceremony as
in neighboring and more important kingdoms. The drum Bitunta, which
Ndahura is said to have given to Rugo, is of particular interest on account of its
minute size; it is little more than three inches high and is covered with the skin
of the monitor lizard.(ll)
First subordinate to Bunyoro, Buzimba came under the suzerainty of Nkore
in the reign of the Mugabe Rwebishengye. Ntare ya Kiboga, during the civil
war with his brother Mukwenda, attacked Nyakairu of Buzimba and plundered
his cattle. Nyakairu, with his family and his brother, another Rugo, fled to
Bunyoro and there he died. During Kabarega's wars, Nduru, Nyakairu's son,
decided to return with his seventy head of cattle. On the way, however, he was
attacked by Nubian soldiers who took all he possessed. Destitute, he arrived at
Ntare's kraal at Muti within the present township of Mbarara, and submitted
to the Mugabe. On payment of one cow which he procured from a friend, he
was restored by Ntare to his kingdom which he continued to rule as the Mugabe's
vassal until the coming of the British.(4)

The Banyampaka who ruled this area were a Bahororo clan which is said to
have come from Rujumbura and to have settled, first near Katwe and then at
Kasenyi on Lake George, under the over-rule of the Babito Bakama of Buson-
gora, who themselves owed allegiance to the Bakama of Bunyoro. With the
valuable salt deposits around Kasenyi under their control, this Hamitic clan
succeeded in the nineteenth century in establishing its rule over a wide area
comprising the much prized grazing land on the southern and eastern shores
of Lake George and on the south bank of the Kazinga channel. This chieftain-
ship is said to have extended south of the channel, from Lake Edward to the
Mpanga river, and to have included the high ground of Bunyaruguru proper.(4)

Although these Banyampaka chiefs can hardly be dignified with the title of
Bakama, they once possessed, so it is said, a drum, Mugonzi, now no longer in
existence. The names of six rulers are remembered; Nkomyo, Ihungo, Goro,
Rutairuka, Kasheshe and Kuri-ofire.(4)
Towards the end of the 18th century, a group of Baganda, known as the
Bakunta, murdered the Kabaka Junju at the instigation of his twin brother
Semakokiro. Fleeing from the vengeance of the ungrateful Semakokiro, they
travelled westward and a portion of them settled in Bunyaruguru.(12) They are
credited with having introduced into that fertile area the banana and the bark-
cloth tree, and their descendants, together with those of Bakonjo settlers of the
next century, form the majority of the mixed population, with its distinctive
dialect, which is found today in Bunyaruguru saza.
In about the year 1890, when Kabarega's army had overrun Busongora,
Kaihura, a nephew of Bwacari the last Mukama of Busongora, fled from the
Banyoro across the Kazinga Channel. In return for a present of seven head of
cattle, Arari one of Kuri-ofire's chiefs allowed him to settle near the present
fishing village of Kazinga. Kaihura was not, however, content to remain an
obscure exile. His cousin Kiboga was the mother of the Mugabe and he offered
to conquer and hold the whole of Kuri-ofire's territory as one of Ntare's chiefs.
To this Ntare agreed, and, with the help of the Banyankore, Kaihura made
himself chief of the land below the Bunyaruguru escarpment which lay between
Lake Edward and the Cambura river.(4)
Meanwhile, Lugard had reinstated Kasagama as Mukama of Toro and both
Busongora and Kuri-ofire's country north of the Kazinga were included in his
kingdom. Kuri-ofire objected, but without avail. William Grant, writing in 1893,
says: "There is a chief called Kuriafiri on an island in the lake who refuses to
recognize Kasagama as Sultan. I told him he would have to do so. Mr. Reddie,9
I understand, made him a free Sultan but as the country was given to Kuriafiri
by Kasagama's people I think it is but right that he should recognize Kasagama
as Sultan."(7) Kuri-ofire's country north of the Kazinga Channel was formally
incorporated in Toro by an appendix to the treaty of 3 March 1894 between
Major Owen and Kasagama;10 and Kuri-ofire withdrew south of the channel
where he managed to maintain his independence of both Kasagama and Ntare
in the high land of Bunyaruguru proper and east of the Cambura until 1900.

When the Mugabe Ntare died in 1895, there was no obvious heir to succeed
him. His nephew, Kahitsi, seized Bagyendanwa and the royal herds, but he was
unable to secure the support of the majority of the Bahima who looked to the
warrior Igumira, another of Ntare's nephews, as their leader. Igumira brought

9 C. S. Reddie (1850-1915). 1890-3 Imperial British East Africa Company service; 1893-5
Uganda Protectorate service; 1895-1914 East Africa Protectorate administrative service.
From June 1892 to March 1893 he toured Busongora and south Bunyoro behind the
Sudanese forts.
10 The treaty, but not the appendix, is printed in Command Paper Africa, No. 7 (1895),
p. 74; see Uganda J. Vol. 13 (1949), 175.

forward a youth, Kahaya, who, he claimed, was the son of Ntare and the right-
ful Mugabe. Matters were further complicated by the appearance of a third
claimant, a boy called Rwakatogoro, the son of Nkuranga, a popular prince
whom Ntare had had murdered some years before in a fit of jealous suspicion.(l)
The following table shows the relationship between the claimants:

Ntare Bacwa Makumbi Nkuranga

? Igumira Kahitsi Rwakatogoro
I ?

Kahitsi, unable to secure the throne unaided, made common cause with
Igumira against Rwakatogoro and appealed to the Protectorate Government and
to the Kabaka of Buganda for support. George Wilson, writing from Kampala to
the Commissioner in December 1895, states that envoys had arrived from
Kahitsi and Igumira asking "for the support of H.M. Government and the King
of Uganda for the son of Mtali". He adds that he had then sent messengers to
Nkore who, on their return, had reported that they had met Kahitsi and Igumira
who had called a conference which "was unanimous in declaring that Mtali had
given birth to a son, by a woman who had died soon after childbirth. This woman
on becoming pregnant, had been sent by Mtali to Gomira's (Igumira's) wife to be
delivered. There were witnesses to the negotiations between Mtali and the parents
of the woman on the occasion of her becoming Mtali's wife."(7)
Rwakatogoro's claim had little support and Kahaya was, in general, accepted
as the rightful Mugabe. Kahitsi who, with Igumira, ruled in Kahaya's name still
hoped, however, to supplant his cousin.
In 1897, the Kabaka Mwanga fled, first to Buddu and then to German territory,
leaving behind a large body of supporters under his Mujasi Gabrielli. For the next
couple of years these freebooters, operating from Kabula, which was still part of
Nkore or Ankole as it shall be called for the remainder of this article, continued
to be a source of trouble to the Protectorate Government, and raided Buddu and
the neighboring country. Though the Government had an active supporter in
a Muhammadan chief in Bukanga called Kahusi, the leading chiefs of Ankole
either sympathized with the rebels or were powerless to resist them. In such
circumstances, it was essential for the Protectorate Government to have effective
control over Ankole. Messengers from Kahitsi and Kahaya had expressed their
wish for the support of the protecting power and, in December 1898, R. J. D.
Macallister arrived to set up a civil station at Mbarara. A military force of one
company was also stationed in the new district.
Macallister had various difficulties to contend with. Kahitsi, who was forced
to hand Ntare's property to Kahaya, threatened rebellion and, like most of

11 Few Banyankore would now agree with the finding of this conference, since it is
generally accepted that Kahaya was the son, not of Ntare, but of Igumira.

the Bahinda chiefs, was jealous of the power of the young Muhororo Nganzi,
Mbaguta. There were still armed bands in Kabula with whom Macallister, with
the small force at his disposal, could not contend. The situation was so difficult
that at one time the evacuation of Ankole was even considered. Since the
enforcement of law and order in Kabula was, inevitably, the responsibility of
the authorities in Buganda, the country was, in September 1899, put under the
control of a Muganda chief with a view to its eventual incorporation in Buganda,
which was formally effected by the Uganda Agreement of 1900. Worse, how-
ever, was to follow. The Buganda Regents managed to convince the Com-
missioner that Ankole, as well as other neighboring territory, had, in previous
years, been tributary to Buganda, and Macallister was informed that an annual
tribute of fifty-five head of cattle, ten goats and eight tusks of ivory should be
paid to Buganda. Macallister replied that no tribute had ever been paid by
Ankole; that the imposition of tribute, following the annexation of Kabula,
would have a disturbing effect; and that, if a tax were necessary, then a hut tax
should be imposed by the Administration.(7)
Macallister and his successor, R. R. Racey, who took over in 1900 were men
of energy and determination. The country was extensively toured and recalci-
trant chiefs were punished. The enthusiasm of these two administrators often
alarmed the authorities in Entebbe who would write despatches (which it would
appear were, as often as not, disregarded) warning the Collectors that they must
not drive "these volatile and excitable natives" too far, lest they should take
themselves and their herds off into German territory, and that, if they could not
administer without force, they must withdraw altogether. A typical rebuke,
administered to Macallister in the form of a semi-official letter of June 1899
from Ternan the Commissioner, reads as follows: "I am sorry to see that you
have had a military expedition in Ankole. Please bear in mind the necessity at
present of keeping quiet. In no case can I send you further troops. If we can't
hold Ankole without a row we must come out of it. F.O. (sc. Foreign Office)
won't have any more rows at present, and there will be trouble if their wishes
are not complied with."(7)
The year prior to the Agreement of 1901 saw the steady expansion of the area
effectively controlled by the Mugabe. Accepting, perhaps too readily, the pre-
tensions which Mbaguta and the other leading Ankole chiefs put forward on the
Mugabe's behalf, Macallister and Racey sought to put them into effect. In this
they were not altogether successful. Despite Macallister's protests that Kita-
gwenda should pay tribute, it was incorporated as a saza in Toro, and neither
Makobore, the head of the Benekirenzi in Rujumbura, nor Rugarama, the head
of the Benekihondwa in Kajara, could be persuaded to accept the Mugabe as his
overlord. Trouble in Bunyaruguru was settled by deposing Kuri-ofire and
replacing him by Kaihura who had previously held only the chieftainship of
Kazinga, and Kuri-ofire's land east of the Cambura was given to Buhweju.(7)
In January 1901, an attempt was made to bring Musinga, the Mukama of
Igara, into Mbarara to acknowledge Kahaya as his ruler. Musinga agreed to
make the journey but custom decreed that Bakama should never meet one
another. Furthermore Musinga feared that he would suffer the same fate as his
enemy Igumira who, having failed completely to reconcile himself to the new

regime, had been exiled from Ankole in the previous year. On reaching the
Kandekye which was the boundary of his kingdom, Musinga produced a hidden
knife and disembowelled himself.(7)
By April, Racey was able to report that Ankole proper consisted of nine
sazas, Isingiro, Nyabushozi, Mitoma, Nshara, Buzimba, Shema, Rwampara,
Igara and Bunyaruguru. Four other areas, he added, had not yet been brought
under the control of the Native Government, Rujumbura, Kajara, Buhweju and
Bukanga. Concerning these, he made the following observations; "Rujumbura
requires special attention. Kajara and Buhweju may come to terms in time and
recognize Kahaya as Principal Chief. Bukanga is independent of Bahima local
or native government by sanction of H.M. Commissioner."(7)
In 1900 a clash had occurred between Kahusi, the Muhammadan chief of
South Bukanga for whom the Administration had the highest regard, and
Kanyabuzana the pagan chief of North Bukanga during which Kahusi lost his
life. Kanyabuzana whom the Banyankore regard as a national hero fled across
the German frontier. The whole of Bukanga was then put under a Muganda
chief, Abdul Effendi, and was for the next seven years independent of Ankole.
With its large population of immigrant Baganda, Bukanga was, during this
period, nearly lost for good to Ankole and absorbed, like Kabula, in Buganda.
The mountainous kingdom of Buhweju was the next problem with which
Racey had to deal. Frightened and suspicious of the Europeans, Ndagara refused
to come to Mbarara or to allow a European to enter his kingdom, though he was
quite prepared to send presents and messages of friendship. Racey was in a
difficult position. On the one hand, he had been warned by the Commissioner
that there must be no more aggression; nor did he wish to risk a second suicide
by forcing Ndagara to come to Mbarara against his will. On the other hand,
Lieutenant G. C. R. Mundy, the officer in charge of the military forces in the
district, who was convinced that failure to enter Buhweju was discrediting the
Administration in the eyes of the Banyankore, was constantly urging Racey to
undertake a military expedition, whilst Mbaguta and the leading Banyankore
chiefs were eager to see their old rival humbled. In April, an armed force which
Mbaguta accompanied entered Buhweju and then withdrew. Ndagara then sent
a relative, Igana, to make blood-brotherhood with Racey, but despite the carry-
ing out of this ceremony, further provocation was given by the Mukama, and in
July Racey again entered Buhweju with the intention of bringing Ndagara to
Mbarara. The expedition and its unhappy climax is tersely described by Racey
in his monthly report for July:
"On the 1st instant a message from Wandagara Chief of Buwenzu was
delivered to me stating that if I again entered his country he would spear me.
I sent him due warning accordingly and entered his country on the morning of
the 12th accompanied by Inst. Wood and 60 constables. On arriving at his
village of Kasungwi, Wandagara did not give himself up but went into an
adjoining shamba with some of his followers where during an attempt to dis-
lodge him he was shot with his son Chiga and a number of supporters who
offered resistance as already reported. The cattle were temporarily removed as
well as a number of women who might otherwise have committed suicide. Chiefs
later came to arrange a settlement and were informed that they must acknow-

ledge Kahia as their principal chief as well as pay a fine of 300 spears. Luwarema
(Ndibarema) the only surviving son of Wandagara was unanimously elected
chief on 25th inst. by all the remaining chiefs of Buhezu who had come in
bringing with them the fine of 300 spears. Luwarema, after acknowledging
Kahia and having handed over the spears, was made chief. The cattle and
women temporarily removed were returned. I am of the opinion that there
will be no further serious trouble in Buhezu."(7)
With the subjugation of Buhweju completed, the time was considered ripe for
a formal agreement, closely modelled on that which had been signed in Toro
in the previous year. This agreement which was signed in August 1901 was a
short and simple document which gave formal recognition to the Mugabe's rule
over the whole of Ankole District, as it exists today, except for the sazas of
Bukanga and Kajara.
In 1907, Abdul Effendi agreed to become one of the Mugabe's chiefs, and
Bukanga returned to Ankole. It was not, however, until some years later that the
western boundary of Ankole was settled. Makobore of Rujumbura steadfastly
refused to acknowledge the Mugabe as his ruler, his ancestors having never, in
fact, been subordinate to Nkore, and, in 1912, Rujumbura was included in the
new district of Kigezi. Rugarama of Kajara was in a difficult position. The
Anglo-German Boundary Commission, 1902-04, which, in accordance with
Anglo-German Agreement of 1890, demarcated the boundary along the parallel
of latitude 1 south, put a large part of his country in German territory. Though
on the whole friendly towards the British Administration, Rugarama found it
hard to make up his mind in which part of his country he would live, and it was
some years before Kajara was effectively administered. Rugarama was, in any
case, an ineffectual chief, and it became necessary first to place a saza chief
over him and then to administer the saza without him through a Muganda.(13)
Finally, in 1914, the boundary was altered in Ankole's favour while the southern
portion of the Benekihondwa territory, together with part of the Bagina lands
of Butaya, was incorporated in Kajara saza.(8) With this frontier adjustment,
Ankole reached its furthest limits and the era of expansion which had lasted
two centuries was over.

(1) Katate, A. G., and L. Kamugungunu (1955). Abagabe b'Ankole. Kampala:
Eagle Press.
(2) Crazzolara, Father J. P. (1950). The Lwoo, Part I. Verona.
(3) Nicolet, Father J. (1953). Mucondozi. Mbarara.
(4) Oral Tradition.
(5) Nganwa, K. K. (1948). Abakozire eby'Okutangaza Omuri Ankole. Nairobi:
English Press.
(6) Williams, F. Lukyn (1935). Early Explorers in Ankole. Uganda J., 2, 196-208.
(7) Records, Secretariat, Entebbe.
(8) Morris, H. F. (1955). The Kingdom of Mpororo. Uganda J., 19, 204-7.
(9) Ndibarema, D. Manuscript in Runyankole on Bakama of Buhweju.

(10) Kanyamunyu, P. K. (1951). The Tradition of the Coming of the Abalisa Clan
to Buhwezu, Ankole. Uganda J., 15, 191-2.
(11) Morris, H. F. (1953). The Balisa Bakama of Buzimba. Uganda J., 17, 71-2.
(12) Oliver, R. (1954). The Baganda and the Bakonjo. Uganda J., 18, 31-3.
(13) Williams, F. Lukyn. Historical Notes in Ankole District Tour Books.
(1946). Nuwa Mbaguta, Nganzi of Ankole. Uganda J., 10, 124-35.
Gray, Sir J. M. (1948). Early Treaties in Uganda, 1888-91. Uganda J., 12,25-42.
Thomas, H. B. (1949). More Early Treaties in Uganda, 1891-96. Uganda J., 13,


TO the casual observer the scenery of Mengo presents a certain monotony
of character: a succession of hills of moderate size rising to an even height
above remarkably similar valleys partly filled by papyrus swamp or high grass;
a repeating pattern of vegetation which over a large part of the district consists
of intermingled grasslands, cultivation and forest. There is a striking absence of
hill-ranges on the one hand, or major valleys or plains on the other. Vegetational
changes and to some extent topographical differences may be noted as one
leaves the environments of Lake Victoria, differences which prompted Baker
(1934) to draw a distinction between the 'Victorian' Plateau and the Interior
Plateau of Uganda as a whole. None the less, although Mengo has less striking
scenery than the neighboring districts of Toro, Bunyoro and Ankole, it presents
subdued contrasts of landscape which are worth considering both in themselves
and as exemplifying the varying effects of the interplay of climate and rock
structure. Such contrasts as occur usually pass almost imperceptibly from one
landscape type to another, but occasionally a sharp boundary can be recognized,
such a one is obvious to the traveller from Hoima, who, after the miles of
monotonous country south of the Kafu River, enters abruptly in the neighbour-
hood of Kiboga2 a group of steep-sided hills with a more diversified vegetation.
At the outset however, it is necessary to decide on what bases one should
establish the broad sub-division of a region such as Mengo. Physiography has
been defined as "a description of nature or natural phenomena in general", and
one might therefore take account of all natural features. The topography,
geology, soil, climate, vegetation and fauna should all be considered in a
physiographical analysis. But since the emphasis has already been laid on
topography, the sub-division of Mengo herein adopted for the purpose of
description will be based on "the structural features of the Earth's surface"-
the more restricted definition of physiography accepted by the British Associa-
tion Geographical Glossary Committee (1951). As emphasized by Frye and
Schoewe (1953) such a sub-division based on fixed physical features easily
observable in the field offers more precision than one based on climate or
distribution patterns of plants and soil types of which the boundaries are less
obvious and tend to be more gradational or in some cases may shift during a
few years.
It is proposed therefore in this paper to describe the general common physical
I A short and slightly modified form of this paper with the title of 'The Physiography
of South Central Uganda' was read at a symposium sponsored by the International
Geographical Union at Makerere College, September 1955.
2 As all localities mentioned could not be indicated on the accompanying sketch map,
Fig. 1, the reader who is unfamiliar with Buganda is advised to refer to Mengo District,
Map reference No. A 1107, Scale 1 inch=5 miles; or Map of Uganda Protectorate,
Scale 1 :1 m., both published by Survey, Land and Mines Department, Uganda.

features of the landscape of Mengo and to draw attention to local variations;
then to discuss the geological and climatic factors which give rise to the present
form of the land, and finally to suggest a physiographic sub-division of Mengo
of which the different units are characterized by a dominant topographical

Throughout the district the amplitude of relief is moderate, with a difference
of about 600 feet between hilltops and valleys; in the northern part of Kyadondo,
Bulemezi, Buruli, Bugerere and Singo this decreases to about 200 feet, while
towards the west in central Singo there is a topographical difference of as much
as 1,100 feet.
Individually, the hills are of two main types.
(a) Flat-topped, plateau-like forms are especially characteristic in the
southern half of the region. The summits of these hills show remarkable coin-
cidence of altitude. This is most marked in an east-west line passing approxi-
mately through Kampala, where all the flat-topped hills are between 4,300 feet
and 4,350 feet with a very few rising to 4,400 feet and occasional flat summits
of small area rising only to 4,250 feet (Fig. 4). In Singo County such hills rise to
4,600 feet and the larger plateaux display gently rolling uplands rising to
5,000 feet. These flat-topped hills were early recognized by Wayland (1921) as
representing relics of a former continuous peneplain. Not only do these hills
rise consistently to a common level, but they have remarkably consistent angles
of slope. Their profiles commonly show a relatively short steep upper slope of
about 24 to 27" inclination, rapidly flattening to long 'pediment' slopes of an
average of 50 to 7 inclination. About 200 such slopes have been measured
mainly in Busuju and Mawokota Counties (Pallister, 1956a) but useful field
work could still be done elsewhere to add to these still scanty observations.
(b) Of less height than the flat-topped hills are rounded, dome-like forms
with evenly inclined or slightly concavo-convex slopes which average 50 to 7
inclination in the central and southern part of the district (Fig. 5), but have a
flatter profile generally in the north, where 3 to 4 are more common.
Occasionally a rounded or hog-backed hill may be surmounted by a short
steeper 'pimple', or in parts of Singo and around Sentema in Busiro the summit
may be marked by a granite 'tor' or 'kopje'. In the highlands of Singo, rather
special conditions give rise to rounded hills with much steeper concavo-convex
The flat-topped and rounded hills have been shown to belong to an evolution-
ary series (Pallister, 1956b). By a process of parallel retreat of hill-slopes, the
area of a hill summit and of the whole hill is reduced without appreciable loss of
height, until the summit is virtually reduced to a narrow ridge or a point; at
which stage it is relatively rapidly reduced in altitude with the formation of a
rounded summit, and then follows a slow process of slope flattening. All stages
can be seen side by side, but in general a locality is characterized by a pre-
dominance of one or other stage in this denudational process-a particular

expression of the 'geographical cycle' as it has long been known from the classi-
cal work of W. M. Davis (1899). Whereas Davis described as 'normal' a con-
tinuous process of summit-lowering, other authors, notably Penck (1924) and
more recently King (1951, 1953) have stressed parallel retreat of slopes. In
Buganda both phenomena are of importance.
The valleys of Mengo show a number of interesting and unusual features.
Although there are a few major rivers, as, for instance, the Nile, Sezibwa,
Mayanja, Nabakazi, and Katonga, these do not occupy broad flat major valleys
as one would expect, but follow moderately straight courses between ranks of
low hills. Correspondingly there are very many tributaries of moderate size,
shorter but often little narrower than the major water-courses, and it is true to
say that in spite of their unimpressive outlines hills are more obvious features
of the landscape than valleys. The valleys themselves, in the central part of the
district display comparatively little flat ground, but consist of the lower
'pediment' slopes of the hills, and the valley-bottom is a relatively narrow zone
of swamp, forest or grassland. Some nearly flat valley-bottoms without streams
are noted below. Furthermore, valleys are generally rectilinear or slightly curved
without the marked sinuosity so common in valleys in regions of similar topo-
graphic relief to that of Mengo. Indeed, marked changes of direction are usually
only brought about by the junction of a tributary with a larger stream. With
local exceptions in Singo and in the east of Kyagwe, valley-side or hill-slope
gullying is rare, and most valleys end in a rounded valley-head or have only
one or two small tributaries. Once again, one is impressed not by the smaller
valley but by the fact that two hills are closer together. The river network over
most of the district displays a mature stage of development with low gradients,
a sluggish movement and courses choked by papyrus. But flood-plains and
meanders which are the usual marks of a mature stream (Lobeck, 1939) are
absent, with the exception of the Kafu River. Yet for swamps, the gradient and
rate of flow of the streams is often unusual. In the scarcity of minor tributaries, it
is obvious that much of the water of the rivers is derived from springs which
emerge frequently little above the average swamp or valley-bottom level. The
rivers are, however, incised in many cases; that is to say, they lie below the
projection of the pediment-slopes on opposing sides, and in crossing a valley
one encounters a relatively sharp increase in slope at the edge of a swamp.
Elsewhere, the rivers have aggraded their beds up to the higher level such that
no break in slope is encountered. In these cases the aggradation has usually
resulted in the silting up of the swamp and little surface water is to be found
throughout most of the year. The swamps themselves are at an earlier stage of
aggradation, being troughs only partially filled by fine alluvial silt. The rivers
have undergone a complex history. Very coarse gravels and boulders at many
localities point to a time or times in the past when torrential floods incised the
valleys, to be followed by successive siltings. The complex Pleistocene history
has been studied in detail in the Kafu valley in the north (Hirst, 1926; Harris,
1944) and in considerable detail in the Kagera valley in south Ankole (Wayland,
1934, 1955).
The course of the Nile on the boundary of Mengo calls for special comment,
although a full description of this remarkable stretch of the river cannot here

be given. It displays two distinctly characteristic reaches. A sluggish lower
course extends from Lake Kyoga up to Namasagali. Apart from the greater
volume of water, this stretch is quite in keeping with the character of other
major rivers such as the Sezibwa. Upstream is a reach of rapids and falls
extending to Lake Victoria. This upper part is obviously very youthful in
character and it seems probable that it developed from a very active tributary of
a major river, the upper part of which is the Kiko River in Busoga. Why this
active tributary was able to reach the accumulating waters of Lake Victoria and
act as the overflow in competition with other streams is a problem as yet
unsolved. It is evident however, that in detail this stretch of the Nile with its
various falls and rapids is controlled by rock structure.
The shore-line of Lake Victoria displays a typically drowned drainage
system. It is described in more detail later. One interesting feature early
remarked by Wayland is the very unusual watershed that separates drainage
into Lake Victoria on the one hand and streams which flow northwards to Lake
Kyoga. This watershed is extremely sinuous and is within a mile or two of the
lake just east of Kampala. A large area of Mengo from about thirty miles north
of the lake-shore lies below the water level of the lake. In other words, Lake
Victoria is perched above much of the land forming northern Mengo. There are
lake benches and gravels bearing witness to the higher levels of Lake Victoria
(Marshall, 1954), and there is evidence in the area of former lower levels of the

In his most influential essay on the geographical cycle, W. M. Davis opens his
discussion with the dictum, "All the varied forms of the lands are dependent
upon-or, as the mathematicians would say, are functions of-three variable
quantities, which may be called structure, process, and time." The inter-rela-
tion of these three variable quantities is the theme of his essay and the foundation
of a great deal of modern geomorphology. The relative importance of these
three factors has since been a matter of debate; King (1953) has placed emphasis
on 'process' as against the more commonly accepted controlling influence of
rock-structure, but his argument has not passed unchallenged (Thompson, 1955).
However, it is a fundamental concept that the present surface of the land is a
resultant of the interplay of atmospheric influences on rock-structures, and since
there is a continual change taking place, the form of the land at any one time is
also a function of age. To understand the present landscape of Mengo, it is
necessary to examine these three 'quantities'.
The principal rock types of Mengo are granitic gneisses and granites, with an
overlying series of metasediments which includes schists, phyllites, quartzites
and amphibolites. There is also a group of moderately flat-bedded sandstones
in the western part of the district. The gneisses and granites are generally fairly
uniform and therefore give rise to little variation in resistance to erosion other
than along joints and fracture planes. By analogy with the present course of the

upper part of the River Nile, it is reasonable to believe that a considerable
number of the water courses of Mengo are controlled by underlying joint
systems, the rectilinear character of which has determined the pattern of the
drainage systems; but the determining structures are now masked by a variable
cover of alluvial deposits. Under humid conditions, granitic rocks are very
liable to chemical decomposition and in most parts of the district these rocks
are now weathered to a considerable depth. The overlying metasediments, by
contrast, are heterogeneous; they are composed of hard, resistant bands of
quartzite and, to a lesser degree, of amphibolite, alternating with soft, easily
eroded schists. Since the series is folded to form narrow elongate bands, it gives
rise to characteristic hills and valleys. The flat-lying sandstones and 'siltstones'
typically found in Singo are resistant to weathering and are responsible for
characteristic landscapes.
Although the laterite 'carapace' so widespread in Mengo is strictly a geologi-
cal structure, its importance and distinction requires some special attention. It
is prominently developed on the summits of the flat-topped hills, and its resist-
ance to normal chemical, and to a less extent, physical break-down has made it
responsible for the characteristic shape of these hills; especially is this so when
it is underlain by badly weathered soft bedrock. Even when broken down into
the loose pisolitic material known as 'murram', it is a controlling factor in the
landscape. Laterite is also extensively developed below the summit levels
(Pallister, 1954, 1956b) and forms protective pavements on many slopes. Its
mode of formation and occurrence elsewhere than in Uganda has given rise to
an extensive literature. Recently D'Hoore (1954) has given some valuable sug-
gestions on the relationship of the hill-top laterites and low-lying accumulations.
Laterite formation in Uganda has not been restricted to one period but has
formed under favourable conditions at any time from mid-Tertiary or earlier
to the present day. The laterite along the shores of Lake Victoria, as seen at
Entebbe, is of late Pleistocene to Recent age.
Without recurrent earth movements of one sort or another, a land surface
would eventually be worn down by processes of erosion to a peneplain, and the
evolution of the landscape would approach stagnation. This is rarely encoun-
tered, and so far as Mengo is concerned there have been earth movements since
mid-Tertiary times-the approximate age of the oldest parts of the present
landscape. One set of movements resulted in a slight buckle or downwarp along
what is now the north-western shores of Lake Victoria. This is evidenced by the
altitude of the old mid-Tertiary land surface. Relics of this surface now form
the flat-topped hills which, in south Kyagwe, as mentioned above, reach con-
sistently to about 4,300 feet, but on the islands of the lake rise only to an
average of 4,120 feet. Other earth-movements which affected the area took
place elsewhere. A new base-level for the streams was early established which
resulted in the dissection and extensive removal of the old mid-Tertiary land-
surface, leaving only the very many scattered relics along a general watershed


00 1 20 25 MILES

% LAKE KY Position and direction.of
view of field sketches
indicated thus:-


6 d

5 0
> 7 -v A

FIG. 1

extending from the Singo Highlands through part of Busuju, Butambala,
Mawokota, Busiro and south Kyagwe. North and south of this broad belt of
country, few relics of the old surface remain. More recent earth movements
along the Western Rift Depression resulted indirectly in a fresh base level being
established for the drainage system and the older valleys were etched to accom-
modate themselves to the new circumstances. Thus arose the incised character
of the drainage. Further complications arose from the influence of regional
tilting sufficient to cause disturbance in the drainage direction in some cases,
as has been described by Wayland. The terraces of the Kafu River in the north
bear evidence of such disturbances, the precise nature of which is not entirely
The present landscape of Mengo is to some extent a legacy of past climates
as well as the result of the present climatic regime. The character of the hill-
slopes with their pediments and the rectilinearity of the drainage pattern suggest
a development under more arid conditions than at present. Coarse pebbles and
boulder beds, seen in places adjoining the Mayanja River (at Miles 6 and 141 on
the Kampala-Mityana road) and at a number of localities in Kyadondo and
Busiro (at Mile 29 on the Kampala-Hoima road) imply the action of strong
torrents in a normally semi-arid region with little protective vegetation. The
extensive development of laterite which would best form under moderately
humid conditions indicates changes from the dry climate. The rise of Lake
Victoria under the probable influence of climate produced the raised sand-
beaches and gently sloping benches in many places.
The equable climate of most of the district maintains a heavy vegetation and
at the present time very little soil erosion is taking place under natural condi-
tions. Climate is responsible for the evolution of landscape by its work of
erosion and denudation, but its influence is very frequently indirect through
the presence or absence of vegetation. The distribution of rainfall is important
since a climatic regime with a marked dry period or periods is most effective;
grass will dry off and burn; the first rains will strike the bare ground, and by
sheet flood and rivulets carry off a considerable amount of fine surface soil.
Later rains will be the less effective as grass grows protectively. Thus a con-
trasting degree and manner of erosion may be found in the lake-shore zone of
nearly equally distributed rainfall, and in the drier zone in the south-west and
northern parts where the dry season becomes more pronounced and relative
humidity is at times low.
Rainfall then has comparatively little direct power to mould the landscape
where the climate sustains perennial vegetation in the form of forest or elephant
grass. Much of the rainfall passes into the ground to assist deep weathering; and
soil movement down slopes is then mainly brought about by means of creep
under the influence of gravity. Thus the pediment slopes may well have been
steepened from the low inclinations common in semi-arid lands, for soil creep
is a slower process than rain wash. Under the protective cover of vegetation,

only the finest materials of the soil reach the swamps, giving rise to the fine silty
character of the valley bottom soils in contrast to the slightly coarser sands
beneath, which must have arrived under somewhat different climatic conditions.

On the basis of characteristic land forms, Mengo may be sub-divided into
the following physiographic units. (See map, Fig. 1.)
1. South-east Lacustrine Belt
The part of south Kyagwe adjoining Lake Victoria from Jinja to Buwuka
Bay is composed of rolling downs of short grass interspersed by deep valleys
often forested and containing free-running streams. Except in the deep bays,
the shore-line is rocky and steep, rising in many places 500 feet from the lake-
shore. The scenery is attractive, and the region has an air of remoteness partly
caused by the barrier of narrow steep ridges which separates it from the rest of
Kyagwe. The population is considerably less dense than most of the lacustrine
zone due to the high proportion of infertile uplands, and possibly to the slow
return of population after the sleeping sickness evacuation early in the century.
In many respects, this region strikingly resembles the Isle of Purbeck in Dorset,
in its short grass downlands, deep valleys, and almost completely land-locked
bays; Nalumuli Bay closely resembles Lulworth Cove (Fig. 2).
Geological structure is a dominant factor in the landscape. The straight coast-
line and parallel ridges are formed of hard quartzites; intervening valleys and
bays on the coast are eroded from weak shales or weathered granite. Grant Bay
is a fine example of the 'arena' structures typical of Ankole-elliptical or circular
depressions some miles across, surrounded by steep-sided hills. The landscape
is comparatively 'youthful' in character. Broad remnants of the old, mid-
Tertiary surface are capped with thick laterite, and valleys are correspondingly
more typical of an upland region than elsewhere in Mengo; the gradients of the
streams are steeper and the amount of valley alluvium is slight. This region
together with part of the Singo Highlands represents a first stage in the destruc-
tion of the old, nearly flat, continuous erosion surface. Indeed on the larger
plateaux which are distinctly dome-shaped, dark soils cover the laterite in places
and may represent the ancient soils of the mid-Tertiary surface.
Rosebery Channel and Buvuma Channel are drowned valleys in structural
troughs and the offshore islands have much the same character as the adjoining
mainland (Fig. 3).
2 and 3. The Eastern and Western Areas of Flat-topped Hills
Central Kyagwe and the counties of Busiro, Mawokota, Butambala and
Busuju with parts of Gomba display the constant repetition of flat-topped hills
with bare laterite 'carapace' and short grass; long pediment slopes thickly
cultivated, and papyrus swamp, with fringing forest (Fig. 4). It is the familiar
landscape of Buganda. The controlling factor of the scenery is the resistant
laterite, underlain by deeply weathered granitic or schistose rocks. The hills
represent a second stage in the reduction of the mid-Tertiary surface. The

FIG. 2
View over part of south Kyagwe, showing extensive remnant of lateritic 'carapace' with forested valleys;
shore line of Lake Victoria is controlled by ridges of quartzite.
[Sketched from an oblique air photograph by J. W. Pallister.

swamps are mostly incised in the valleys and are perennially wet. The water-
shed between Lake Victoria and Lake Kyoga to the north forms no marked
feature, the head-waters of the two drainage systems inter-digitate, or through-
valleys exist where the direction of flow is not obvious. These valleys have
many interesting features. At the point where the Entebbe-Kampala road
crosses the railway near Katwe village, four streams take their source; three
flowing into Lake Victoria at Port Bell, at Gaba, and near Kazi, and the fourth
marks the headwaters of the Mayanja River flowing eventually to the Kafu and
thence to Lake Kyoga. The watershed is only 140 feet above the present level
of Lake Victoria. Two other tributaries of the *Mayanja head into streams
flowing into the lake and these similarly rise only to about 150 feet above the
lake. Moreover, the watershed region of one, just south of Budo College, has
sand and alluvium at its crest, hence the true valley floor must be less than
150 feet above the lake. An interesting study could be made of the history of
this anomalous drainage area. It is sufficient here to point out that Lake Victoria
could only rise for about 150 feet before it would leak out of more than one
valley on this north-west shore.3 High-level lake-benches recorded at Entebbe
are listed by Marshall (1954), the upper two of which are at 175 and 250 feet
respectively above lake-level. In consideration of the low watersheds just
described it is doubtful whether the abrupt change of slope on Nsamuzi Hill
at 3,970 feet can be regarded as marking an old lake-level, nor can the various
flats at 175 feet above the lake be regarded with certainty as lake-benches. At
neither levels have gravels been found.
The vegetation of the swamps is in delicate equilibrium and slight changes in
the depth of water may give rise to new ecological assemblages. The vegetation
zones of lake-shore swamps have been described by Eggeling (1934).
The soils of these two areas display a catena complex and are derived largely
from the break-down of the laterite and probably only indirectly from underly-
ing bedrock.
4. The Mabira Forest
The eastern part of Kyagwe, largely covered by the Mabira Forest, is an
easterly extension of area 2 but it has some distinctive physical features. There is
a considerable area of rolling high ground which may owe its character to the
abundance of laterite derived from widespread iron-rich amphibolite, and many
streams are free-flowing, contrasting therefore with the Sezibwa River and its
main tributaries immediately to the west. Moreover, the Nile, forming the
eastern boundary passes over falls and rapids and has a quite unusual character.
Nevertheless, it is rather the continuous forest cover which makes this area
distinctive. It has been well described by Pitman (1934) in the first paper in the
first issue of this Journal.
5. The Central Area
Central Mengo displays a number of intermediate characteristics. It possesses
few flat-topped hills, the common form has a broad, rounded outline (Fig. 5).
Many of the valleys are dry, grass-filled or only intermittently swamp; papyrus
3 According to Hurst (1952) a rise of only about seventy feet would cause Lake Victoria
to spill over into central Tanganyika.

FIG. 3
Islands in Lake Victoria displaying the characteristic shapes of mainland hills but at lower altitudes owing to
structural down-warp.

FIG. 4
Typical flat-topped hills of Busuju and Butambala Counties.
[Field sketches by J. W. Pallister,

fills larger courses but only occasionally smaller streams. This region represents
a third stage in the erosion cycle. It has a distinct uniformity but the boundaries
are difficult to demarcate as the type of country merges into higher hills south-
wards and into lower ground of less relief northwards. However, limits may be
set by physical features, and along part of the northern boundary a line of
quartzite hills stretching from the Mayanja River eastwards through Bukalasa
to Nakifuma marks a distinct geographical break, evident to anyone travelling
on the Kampala-Nakasongola or Kampala-Bugerere roads. A few steep-sided
narrow ridges, often forested, strike across the country running almost east-west.
These are formed of very resistant, brecciated, quartzose rock. This direction is
followed by the straight stretches of a number of the streams, particularly
around Namulonge, and is an indication of some structural control (Hepworth,
Laterite is a very common feature of the surface of the ground, but it is
generally reduced to 'murram'; only very occasionally is a pavement of massive
laterite developed.
Land usage is here somewhat different from the areas to the south, for cultiva-
tion can be carried over the hill summits with the improvement of soil on the
breakdown of the laterite. Grazing is possible in the dry aggraded valley-

6. The Northern Lowlands
The northern half of Mengo is an undulating plain of low relief, with numer-
ous sinuous watercourses which are shallow and dry for most of the year. They
are silted up with fine sand and in places especially along the Kafu and Lake
Kyoga, the landscape approximates to an alluvial plain studded with numerous
low rounded hills. The regularity of these hills like inverted saucers is well seen
on air photographs. Soils are not characterized by much 'murram' but pave-
ments of laterite flank many of the slopes at a short distance below the surface.
Here we have a fourth stage of the erosional cycle representing the approach to
another peneplain.
Seasons here have more contrast and the movement of soil towards lower
levels may be accelerated by sheet wash in heavy rains when the grass is burnt.
The vegetation is of the savannah type; climate and topography contrast with
that of south Mengo.

7. The Singo Highlands
As mentioned earlier Singo possesses plateaux rising to 5,000 feet, for which
a group of flat-lying sandstones is responsible. These rocks have the same
influence as the plateau-forming laterite but to a greater degree. The drainage
has bitten deeply into the plateaux to form steep-sided valleys often over
1,000 feet deep. The slopes of the valleys are formed of moderately resistant
phyllites which, once the sandstones have been eroded away, form the distinct,
steeply rounded outlines typical of Ankole. Elsewhere granite forms an upland
with tors or small kopjes diversifying the landscape. To a great extent, the stage
of the erosion cycle in this region is less important in moulding the surface of
the land than rock-type and structure (Figs. 7 and 9).

FIG. 5
Rounded hills of Kyadondo in central Mengo.

Katera 4600

Kala9aaI- A560

FIG. 6
The Northern Lowlands from the edge of Singo Plateau near Katera, looking eastwards; the valley of the
Kitwe River cuts through a ridge of resistant quartzite.
[Field sketches by J. W. Pallister.

8. Lake Wamala and the Katonga Valley
Whereas previous sub-divisions of Mengo just described owe their character-
istic landscape to denudational forms-the stages of erosion or characteristics
assumed by different rock-types, the environment of Lake Wamala is marked
by the recent accumulation of alluvial sediment in the broad drainage basin of
the Kibimba River and its tributaries. In addition, the flat-lying or gently
folded Singo Series of rocks yields gentle slopes which dip under the flat
alluvium. The silting up of Lake Wamala is rapidly progressing; maps based
on surveys in 1908-9 show a considerable area of open water which is now
practically all papyrus swamp; and the extent of grassland which is only
occasionally flooded has encroached on the lake. Myanzi Railway Station
occupies land marked as swamp on early maps.
The Katonga River is flanked by terraces of sand and silt which indicate a
complex history and the northern tributaries are partially dammed by sand-
banks owing their origin to higher levels of Lake Victoria (Johnson, 1954). The
broad flats at its mouth are deltaic deposits formed when the lake was at twenty-
five feet above its present level.

9. The South-west Borderland
Between the Lake Wamala depression and the Nabakazi River is a slightly
undulating plain dissected sharply by the drainage (Figs. 8 and 9). The rocks are
not well exposed but where occasional hills rise above the general level, they are
usually made up of rocks of the Singo Series. The flat watersheds are therefore
believed to represent a fossil erosion surface on which, in Pre-Cambrian times,
the Singo Series was laid down probably in a shallow inland sea. The controlling
factor of the landscape is therefore once again a peneplain but one distinct from
that which forms the characteristics of other parts of Mengo. The rocks are
such that streams have cut back rapidly in steepsided valleys-well seen along
the road between Kasanda and Mubende (Fig. 9).

Variations in the land forms and scenery of Mengo are not striking, but they
are sufficient with which to sub-divide the district into physiographic units.
Each of these has a general geographical character. Each sub-division is found
to be predominantly controlled by the structure of the bed rocks, or by the stage
of erosion of an earlier peneplain, or by accumulation of sediment. They
illustrate strikingly the varying interplay of structure, process and stage of
development of the geographical cycle. It may be found possible to use these
physiographic divisions as a basis for soil, vegetation and occupational studies.

Grateful acknowledgement is made to the Director of the Geological Survey
of Uganda for permission to publish this paper, and thanks are accorded to the
writer's colleagues whose work of geological mapping in Mengo has been of
great assistance.

Butal PLateau .Bul.o (525o) Bik.o(528)
Lusungw" (4395) MpLnve (olo) Makokot (4830) Buugoga (52o)
MPL VeB(Solo) (.4Soo)

FIG. 7
The rolling Singo Plateau on the right and Buta Plateau on the left seen from Luzira Hill in Singo, looking about
north: the foreground is composed of an irregular granite upland surface.

Lubona (4529) KaKL at (4472) NSroroo
tYaUa (4350)

FIG. 8
The incised plain of western Mengo, sloping down on the right towards Lake Wamala; view from near the
Nabakazi River looking east-north-east.

FIG. 9
The high Buta Plateau and incised plain of western Mengo; the valley of the upper part of the Nabakazi in middle
[Field sketches by J. W. Pallister.

Baker, S. J. K. (1934). The Population Map of Uganda: A Geographical Interpreta-
tion. Uganda J., 1, 134-44.
British Association Geographical Glossary (1951). Geogr. J., 117, 458-9.
Davis, W. M. (1899). The Geographical Cycle. Geogr. J., 14, 481-584. (Reprinted in
Geographical Essays. W. M. Davis, ed. D. W. Johnson, 1954.)
D'Hoore, J. (1954). L'Accumulation des Sesquioxides libres dans les Sols tropicaux.
Pub. de I'.N.E.A.C. Ser. Sc. No. 62, p. 132.
Eggeling, W. J. (1934). Notes on the Flora and Fauna of a Uganda Swamp. Uganda J.,
1, 51-60.
Frye, J. C., and Schoewe, W. H. (1953). The Basis for Physiographic Subdivision of
Kansas. Trans. Kansas Acad. Sc., 56, 2, pp. 246-52.
Harris, N. (1944). Ann. Rept. Geol. Surv. Uganda for 1943, pp. 22-9.
Hepworth, J. V. (1956). Aspects of the Geology of the Bombo Area, Mengo, Buganda.
Rec. Geol. Surv. Uganda for 1954.
Hirst, T. (1926). Unpub. Prog. Repts. Geol. Surv. Uganda.
Hurst, H. E. (1952). The Nile, London.
Johnson, R. J. (1955). Lake Levels and the River Katonga. Unpub. Rept. Geol. Surv.
King, Lester (1951). South African Scenery, Edinburgh.
(1953). Canons of Landscape Evolution. Bul. Geol. Soc. Am. 64.
pp. 721-52.
Lobeck, A. K. (1939). Geomorphology, New York.
Marshall, K. (1954). The Prehistory of Entebbe Peninsula. Uganda J., 18, 44-57.
Pallister, J. W. (1954). Erosion Levels and Laterite in Buganda Province, Uganda.
C. R. XIXth. Int. Geol. Cong. Algiers. fasc. 21, pp. 193-9.
- (1956a). Slope Development in Buganda. Geogr. J., 122, 80-7.
- (1956b). Slope Form and Erosion Surfaces. Geol. Mag. (In the press.)
Penck, W. (1924). Die Morphologische Analyse trans. as Morphological Analysis of
Land Forms, by Czech, H., and Boswell, K. C., London. 1953.
Pitman, C. R. S. (1934). The Mabira Forest. Uganda J., 1, 7-16.
Thompson, H. D. (1955). Canons of Landscape Evolution-A Discussion. Bul. Geol.
Soc. Am. 66, pp. 1205-06.
Wayland, E. J. (1921). Ann. Rept. Geol. Surv. Uganda, para. 49.
- (1934). Rifts, Rivers, Rains and Early Man in Uganda. Journ. Roy. Anthrop.
Inst., 64, pp. 333-52.
- (1955). The Pleistocene Geology and Prehistory of Uganda, Pt. 1. Pleistocene
Geology. Mem. No. VII, Geol. Surv. Uganda (in preparation).

M Y purpose here is to describe, not so much the Karamojong and Suk tribes,
as the relations, unfortunately bitter, which have existed between them
for as far back as we can trace, that is for about 25-30 years, when a large
section of the Suk in search of pasture for their starved cattle overflowed from
the West Suk and Baringo areas of Kenya and settled in the south-east corner
of Karamoja District. There they were given their own chiefs and administered
by the Uganda Protectorate Government based at Moroto. Their lot is little
better there than it is in Kenya since the hot, eroding winds from Turkana and
overstocking of cattle have all but ruined the land and turned it into a semi-arid
waste. If it is to be saved, drastic and urgent measures are required, and these
have been put in motion, by dispersing the concentrations of cattle in the vicinity
of water-holes, by increasing the boreholes and pumps, by buying and exporting
to other parts of Uganda as many cattle as these cattle-loving tribesmen will
part with and by instituting a system of reserve grazing areas on a rotational
basis. The problem, however, is fraught by numerous complications and may
prove insoluble. On account of the poor soil, cultivation is little practised and
the staple diet is meat, blood, milk and imported posho (a kind of maize flour
made into porridge).
Karamojong and Suk men go naked, but the women wear cow-hide skirts,
plain for the unmarried girls and gusseted for the married women. Red, blue
and yellow is the favourite colour scheme for the beads worn by Suk men.
A mud skull-cap with red pompons and thorn-resisting giraffe-hide sandals
make up the male attire, which is rounded off with the inevitable spear, a well-
made murderous looking weapon eight feet long. Some of the younger Suk carry
bows and arrows, in the use of which they are extremely skilled; poisoned
arrows are often used on game. Small of stature compared with the Karamojong
they are a courageous tribe and have, on various occasions, put the numerically
superior Karamojong to flight, inflicting heavy casualties.
The Suk are, generally speaking, stubborn, cunning and obstinate, their
language is bewilderingly complicated to learn and they are extremely difficult
people to understand. It was from their midst that the quasi-religious and
proscribed society, known as the Dini ya Misambwa sprang.' This society was
responsible for the clash near Lake Baringo in 1950 when many lives were lost
including those of a Kenya Government District Officer and two European
Police Officers, who were speared to death by fanatical members of the sect.
However, in Karamoja, from a political point of view, the Suk have given
little cause for concern. Life in the Suk part of south-east Karamoja is a battle
for existence, the erosion and general aridity in the dry season have to be seen
to be believed and in bad years Suk cattle have died in their hundreds of starva-
tion. This is a major tragedy to the Suk, as their lives are centred round cattle,
they seldom think long of anything else, and their love of cattle, which
I Dini ya Misambwa by the Rt. Rev. L. C. Usher-Wilson, Uganda J., 16 (1952), 125.

transcends their love of the family, is akin to worship. It is a fact that when
going into battle the Suk warrior, to give himself courage, shouts out, not the
name of his wife or sweetheart, but the name of the big bull of his herd. To the
Suk (and the Karamojong) the bull represents everything that is strong, virile,
courageous and masterful, and the Suk have a repertoire of amusing songs in
which the bulls and cows of the herds are likened to the men and women-not
only metaphorically but physically; reproductive organs come in for a great
deal of lewd comparison!
The Karamojong as a tribe need less description as they are so much better
known, at least in Uganda. There is evidence that before the overflow of the
Suk into Karamoja, Karamojong lived as far south as the area between Kacheliba
and Kapenguria and some of the villages near the road linking these two places
bear Karamojong names. Suk names are quite different and do not resemble
The naked Karamojong are fine physical specimens. Six feet is a common
height, and, in their ceremonial leopard-skins and ostrich-feathers, grasping
their long spears, they are a fearsome sight. Their love of cattle is no less intense
than that of the Suk. Bride price among both tribes is high and paid in cattle, the
average being anything from fifty to a hundred head; prices of cattle being
roughly from Shs. 150 to Shs. 200 per head, brides in Karamoja are therefore
expensive items. Because of this custom and the ancient one of spear-blooding,
we have two of the fundamental reasons for inter-tribal cattle raiding and its
accompanying bloodshed and misery.
Spear-blooding is by the lore of both tribes necessary when a young Suk or
Karamojong has passed through what is known as the 'initiate' stage and
desires to get married, and has first to prove to the elders of the tribe that he
is a man. In early times he had to set out alone, armed only with a spear, and
hunt and kill single-handed one of the lions or elephant which then roamed
the southern Karamoja plains. The young man then reported to his elders at a
formal baraza (meeting), described what he had done and proved it by showing
the blood on his spear blade and by producing the animal's tail. Having passed
his test his only remaining problem was to find sufficient cattle to pay the bride-
price, which, in those times, was not high and consequently did not present
much difficulty to the average young Suk or Karamojong with cattle-owning
relations from whom he could borrow or beg. The balance was often made up
by minor sneak-thieving from neighboring herds.
Nowadays, however, there are virtually no big game animals in southern
Karamoja and the very high bride-price leaves the tribesmen only one choice-
large scale cattle-raiding of the neighbours, Suk or Karamojong as the case may
be, or of the mild Sebei in the south across the border in the Elgon area. The
Sebei invariably take these murderous raids (in many of which people are
killed) lying down and seldom retaliate.
The land in the south of Karamoja is covered thickly with the short acacia
and the vicious 'wait-a-bit' (or hooked) thorn abounds. Attempting to run
through such country results in painful and mildly poisonous scratches and
badly torn clothing. The Suk and Karamojong however, naked as they are, run
full tilt in a straight line with stolen cattle and never seem to be any the worse.

The pursuers' difficulties are increased by heavy rifles, ammunition and boots,
and it is obvious that they stand no chance of catching up with the raiders.
Good radio communications and interception are therefore the best practical
On one occasion, not long ago, during a series of raids and counter raids, a
small party of two policemen and a chief managed to intercept a raiding party
which turned out to be much larger (about 300 strong) than they had expected.
As they confronted each other at a distance of about 70 feet and whilst the
chief was advising the gang to leave the stolen cattle and go home, a warrior
was seen creeping up on each flank, spear poised ready to throw. Both policemen
took careful aim and shot the two warriors down before they could throw their
spears, quickly reloaded and awaited the inevitable rush, feeling no doubt that
their last moment had come. It should by rights have come-three men against
300-but for some extraordinary reason the gang turned tail and fled. The three
guardians of law and order were very lucky men that day. Not long afterwards
they were presented with medals by H.M. the Queen when she visited Uganda.
Men like these are to be admired-unflinching courage and coolness in such a
terrifying situation is a source of inspiration. Good luck to the constables-
they are corporals now.
Karamojong and Suk invariably refer to each other as 'the enemy' and one
always knows to whom they are alluding. I have often spoken to them of this
bitter enmity between them. The Karamojong say that they themselves are
peace-loving people and are perfectly happy to graze their herds and mind their
own business, but how can they do so when, without any suggestion of provoca-
tion they are attacked and robbed by those unspeakable Suk, who are nothing
but troublemakers and should be chased back into Kenya?
Cross over the border of Pian and Upe into Suk country, talk to the elders
and hear them, also with perfectly straight faces, say that they themselves are
peaceful people who abhor the very thought of cattle-raiding. All they want is to
be left in peace to graze their stock, but how can they when, for no reason at all
they are victimized by those cursed Karamojong, who are scourges on the face
of the earth and ought to be put down?
By far the more polished presentation of woe and self-righteousness comes
from the Suk. The average Suk is a raconteur of no mean ability and I think
could hold his own with any Arab. The most barefaced lies are expertly dressed
and disguised in such impressive finery that, unless one knows them fairly well
one is inclined on first impressions to feel oneself becoming rapidly very sympa-
thetic and sorry for them and wondering why on earth something more substan-
tial has not been done long ago to better their lot, poor fellows!
It is a known fact that in the manyattas (homesteads) lying near the better
known Suk raiding routes, children are schooled to answer any strangers'
questions by the words "I don't know". Even when asked "What is your father's
name?" or "Do you live here?" the answer is the same, and anyone trying to
find out about cattle has an extremely difficult task on his hands. It would be
unfair to say that this practice exists only among the Suk, as I have witnessed it
in Pian country where one of the better-known clans of the Karamojong live.
Tracing stolen livestock obviously has its difficulties, but a knowledge of

tribal customs is a tremendous asset in revealing otherwise closed avenues of
exploration. As an example of this, a raid took place not long ago in which one
of the raiders was killed. The others got away safely with the cattle which were
soon dispersed. The defenders, as usual, could not identify the dead raider as
there is seldom any social intercourse between them, and the case looked a
difficult one to clear up. However, a close inspection of the dead raider's head-
dress showed that a part of it had been dyed with a particular kind of vegetable
dye which is normally only used by a very small clan who sympathize with the
Suk and who live on Mount Kadam in south-east Karamoja, west of Amudat.
The Mount Kadam area was then worked on, and as usual no one admitted
having seen any stolen cattle (all expressed horror at the very suggestion!).
However, in the course of routine inquiries a married woman was noticed to be
wearing a thin strip of goat skin round her leg, just below the knee, with the
white hair showing. That was enough, as Suk custom requires a widow to wear
just such an ornament for one month after the death of her husband. Interroga-
tion of the woman eventually revealed that the dead raider was in fact her
husband, whose friends were quickly rounded up before they could decamp and
the stolen cattle were recovered soon afterwards.
In some sections of the Suk tribe, a woman shaves her head within a few days
of her husband's death, and by knowing these customs, which are rigidly
adhered to, much can be discovered.
Karamojong and Suk have many of the same traditions and customs. Blood-
drinking is common to both and is a most impressive (and somewhat nauseat-
ing!) ceremony, particularly when one is camping in the bush and witnesses the
spectacle by moonlight or the flickering light of the camp-fire. Firstly, a good
fire is lit and the bull dragged into the clearing and held by one or two tribesmen
by a thong tied to a hind leg. A spear is then driven-they are sometimes thrown
a very short distance-without further ado through the bull's heart; it is said
to be unlucky if the heart is missed the first time. As soon as the bull falls to the
ground and whilst it is still alive but bleeding to death, the chest cavity is quickly
cut wide open with spears.
The men have by this time already taken up positions close to the bull; elders
first in order of seniority and the lesser fry lining up behind for any blood that
may remain. One by one they kneel down and put their heads deep into the bull's
chest cavity and drink the hot blood as it gushes out. Standing up they move
aside to let others drink, and present a fantastic sight with blood running freely
down their chins and chests, grinning broadly meantime. This completed, the
bull is dragged whole on to the fire and then removed again in a few minutes as
soon as the hair has been singed.
An elder or witch doctor then removes the bull's intestines, takes the dung
from them and hands it round to those present, a small portion to each man.
They smear it on their chests as insurance against bad grazing and non-produc-
tive women. Whilst this is going on the elder or witch doctor daubs dung on
small stones on the ground in a variation of patterns and, mumbling incantations
and formulae over them, attempts to discover what the future holds in store
for the tribe. This practice is widely used by both tribes in southern Karamoja
when they are seeking to discover whether the signs are favourable to success

in cattle raids. Details vary, but the basic technique never does, although I
believe that the Suk of Loro area north of Amudat and in the Karasuk hills
east of Loro practise an additional system of tossing a pair of giraffe hide sandals
in much the same way as 'Two-up' is played in Australia with two pennies. If
the sandals fall in mutually complementary positions a certain number of times
in succession, the signs are favourable and the eating of the bull which follows
immediately afterwards assumes a sinister character; normally, however, a
raiding feast does not just happen, but is arranged beforehand.
Should the signs be favourable, the bull's carcase is cut up with spears (which
are invariably used, never knives or choppers) and distributed in small portions:
most portions have special significance and are handed out in accordance with
social status. Excitement mounts and, whilst the men are devouring the raw
meat, songs are sung and women in the background keep up a set rhythm by
stamping and clapping. This part of the ceremony is likely to go on for hours,
the men under thirty beginning to dance and vie with each other how high they
can jump. This they do to quite an impressive height, each man jumping singly
or in small sets in time to the communal stamping, clapping and chanting; he
faces a young woman and they bob up and down alternately. No musical instru-
ments or drums are heard, only the monotonous honking of a long horn made of
tough hide, which blares out a single note in time to the stamping. The songs
are often about the prowess of the men and their courage and daring in battle,
about the big bulls of the herd and how the men resemble them in strength. No
man would think of risking the women's ridicule by refusing to go on a raid;
life would be unbearable then.
All this chanting and jumping works the younger men up into a pitch of
excitement, when all thoughts turn in the obvious direction-a raid. One of the
more prominent young bloods is given the bull's testicles to eat. This is a signal
honour and has the effect of transferring the bull's power and fighting spirit to
the eater, who will, no doubt, be one of the leaders of the raid.
On more peaceful occasions this is not done. For instance, at a feast to
celebrate the opening of dry-weather grazing grounds the dance ends usually
with the older people making their way home, whilst some of the younger men
go off into the bush with the girls with whom they have danced. Quite a lot of
native beer (a thick, porridgy substance made from millet) is consumed on these
occasions, but drunkenness as encountered in other parts of Uganda is almost
unknown. Alcoholic spirits are quite unknown to both the Suk and the Kara-
However, this particular feast is a serious one. The bull has been eaten, the
signs have been found favourable, one of the leading warriors has partaken of
the most honoured morsel, the older people and the women have gone home,
the younger men sit down and a council of war begins. The sounds of revelry
have subsided and the men, gathered round the dying embers, know there is
serious work afoot. As raiding feasts invariably take place in isolated parts of
the bush, interruption by police or chiefs is rare.
This is an authentic account of the preparation for a raid gleaned from
descriptions by both Karamojong and Suk who have actually been members of
a raiding party.

Raiding is by no means haphazard; there is nothing hit or miss about it, and
the skill and technique displayed by both Suk and Karamojong compare very
favourably with the activities of the Commandos in the late war. For speed,
ferocity, cunning and first-class fieldcraft there are few operations of modern
armies from which Suk and Karamojong could learn very much-they are
experts in the art.
Firstly, there are always, in troubled times, regular day and night recon-
naissance movements by tribesmen in twos and threes, ostensibly wandering
about looking for honey or new water-holes, but in fact subjecting the move-
ment of enemy cattle to a very close watch. Both tribes are partly nomadic, the
Suk rather more so than the Karamojong. There is always to hand, therefore,
up-to-date and accurate information of dispositions of enemy cattle, and the
briefing for a cattle raid usually commences with an exchange of such informa-
tion, with a view to choosing the most profitable and practicable target.
The objective having been agreed upon, the leaders lay down a plan of action.
Having done this sort of thing many times before, they know exactly how to
organize the operation. Leaders are usually rather older or middle-aged men
who own large herds (or, to be more accurate, have acquired many of them by
questionable methods) and whose social status and authority are therefore high.
The gathering is first split up into two main groups; the attack group, by far
the larger, and the reception group. The reception group, which might number
about ten or fifteen only, does not wait for the main briefing-it knows the
place and approximate time of the intended raid and the withdrawal route, and
so it departs each man going his separate way to the area of a small community
where he enlists reception groups and instructs the local people in the disposal
of the stolen cattle which are intended to be driven into that particular area.
Every man has now acquired two spears and a shield (made of toughened
giraffe hide and fairly effective against spears and arrows) and the attack group
is divided into smaller parties of perhaps ten or twenty which correspond to the
number of herds to be stolen. A herd usually comprises about fifty head of
cattle and as many as ten herds are kraaled together in the bomas at night. This
attack party then moves off quietly for perhaps ten or fifteen miles along bush
paths, maintaining silence and making very good speed. At about a mile or so
from the objective the party halts. It is then usually well after midnight and
scouts, perhaps three or four men, make a final reconnaissance.
An hour or so later they return with exact details of the kraals, the cattle
inside them and how many enemy tribesmen are sleeping in and around the
More waiting; then, as the darkness begins almost imperceptibly to thin,
everyone moves forward very carefully and quite noiselessly. Each man has a
spear and shield in his left hand and the other spear, the one which will be used
first, in his right. In single file, crouching low and edging forward, the party
gradually surrounds the kraal. Nothing is heard but the lowing of some of the
cattle inside. Then more waiting.
Suddenly comes the sound of a hyena-a common sound at night in Kara-
moja. Immediately the whole party rises; one or two of the tribesmen sleeping
outside are stirring as they sense danger. The raiders are in groups facing the

kraals and are now rapidly moving forward. Having previously spotted the
herdsmen asleep they are on them in a flash, and a few muffled groans signify
that spears have gone home. The kraals are now open and the raiders inside.
The cattle rush out, bellowing in fear and pain as spears are jabbed into them
to hurry them up. Dogs are barking and people are stirring in their huts. But
raiders are also at doorways, thrusting and stabbing furiously in the darkness.
There is confusion everywhere-noise, dust, snarling dogs, crying babies and
bellowing cattle.
Next moment there is nothing-raiders and cattle have disappeared into the
bush and all that remains is the moaning and wailing of the wounded and the
yelping of badly frightened dogs. About 300 cattle have gone in a twinkling and
the air is soon filled with the sound of the alarm-the survivors send out
warnings by means of clapping their hands intermittently over their mouths,
making a kind of yodelling scream at the same time, and also beating their
cooking pots. These strange sounds carry far, neighboring settlements are
quickly on the alert and word is soon on the way to the nearest police outpost-
probably a distance of some five to ten miles.
But the damage is done and the raiders are well away, swiftly and unerringly
towards their county border. They are highly pleased with the night's work and
sing exultantly-although somewhat imprudently-as they speed along, boast-
ing of their prowess and courage and thinking what a figure they will cut with
the women.
However, this has happened many times before and the enemy alarm system
is good. Before the raiders reach their borders they cease their singing and begin
casting anxious glances behind them, spurring on the herd to greater speed. The
first reception parties show up and the peeling off process begins, small groups
of stolen cattle being driven off from the main herd at a fast rate into the bush,
there to be further disintegrated and mixed with local herds.
The anxious glances behind are soon explained as one of the raiders in the
rear falls to the ground; the men nearest him see an arrow sticking deep into
his back as they press on with greater speed. Down goes another raider and the
8 feet long spear deep in his chest is plain for all to see. He was on a flank and
made an easy target for the enemy tribesmen who, in their scores, rush on the
raiders from both sides. Battle is joined-dawn is breaking as spears and
arrows fly thick and fast. The raiders have been careless and over-confident;
instead of leaving the stolen cattle, grouping together and meeting the attack
as a body, they greedily held on to them too long, and have been split into two
almost disorganized groups. At least half-a-dozen warriors are now on the
ground and the attackers gaining the upper hand, when over the brow of the
hill comes a rescue party in strength-they are the main reception parties, who,
having summed up the situation, have banded together and are sweeping in
with thrusting, lunging spears. Before long more fighters are dying on the
ground and what is now left of the retaliation party runs back in confusion,
disorganized and beaten, some of them with gaping spear wounds. The stolen
herd is quickly and expertly re-assembled and the raiders are off once more.
Soon they are back in their home county, the cattle have been taken away in
twos and threes and hidden, and the men are glad to get home, hide their blood-

stained spears, drink long and deeply of the calabash of fresh milk and no doubt
wonder how many of their comrades have fallen.
It is hard to imagine these warriors surviving the major injuries they receive
in these battles-not only surviving them but actually walking home in such a
condition. I saw a young Suk after a fight; he had seen a spear coming and
raised his shield to protect his face; the spear glanced off the shield and went in
just under his chin, the point protruding between his eyes and slightly above
them. His whole face came off as though on a hinge, held together by flesh and
skin on the left side of his face. He crawled away into the bushes and hid until
the battle was over. He then put his face in position with one hand, tied three
long strips of grass round it to hold it on, walked home 15 miles and lived to tell
the tale!
However, word of the attack has by now been sent to the nearest police outpost
of three men. Two of them with chiefs and their police and trackers are follow-
ing the stolen cattle, whilst the other one is making all speed on his bicycle to
warn the nearest police outpost equipped with radio, so that word can be sent to
Moroto, some fifty miles away from where armed reinforcements will soon be
The raiders, however, now being home, are not unduly worried. The first
things to do are to bury the shield and second spear and to rub the other spear
vigorously with sand-an excellent abrasive which cleans off bloodstains most
effectively. Those of the raiders who are injured in any way cannot stay at home
unless they live in the most isolated parts far distant from chiefs and police, as
they will soon be visited when awkward questions will be asked about those
injuries. So, having rested for an hour or two, off they go to uncles or cousins in
another county, there to lie low until the hue and cry has abated.
The chiefs and police have by now arrived at the scene of the battle. There is
little to do there beyond chasing away the already gorging hyenas and carrion
birds and attempting to identify the dead. On they go as the trail is still warm
and no time must be lost; it is raining heavily and darkness is not far off. They
are many miles out in the bush and a lion on the prowl is roaring nearby. It is
quite dark when they arrive at the place where the cattle were split up and
before long the trail has dwindled to nothing. They have done their best and at
least ascertained through what particular areas the stolen cattle have passed.
This information may be of the utmost importance if the inhabitants of those
areas (who could hardly but be aware of the noisy passage in daylight of a large
number of cattle close to their huts) deny knowledge of the cattle and the imposi-
tion of a communal fine is considered.
The reinforcements from Moroto arrive at dawn and one of the first tasks is
to collect the owners of the stolen cattle and take them over into the raiders'
county to help search for and identify their stock, as, without the owners or
herdsmen, searching is a hopeless task when no one could identify any of the
cattle even if they were found. Owners and herdsmen, fearing for their lives in
enemy territory, invariably decline to go, intending no doubt to enlist in the
next retaliatory raid and make good their losses that way. This refusal by
owners and herdsmen is a major obstacle to the success of the investigation.
However, quite a reasonable number of cattle are picked up in twos and threes

in widely scattered places, often on information from raiders or their friends
who have quarrelled over the spoils or from wounded raiders who have not
been quick enough at getting away from their homes before the chiefs and
police arrive. But by this time another retaliatory raid has been carried out
(they follow swiftly and no time is lost) and cattle in the hundreds are hope-
lessly mixed up on both sides of the tribal borders.
Evidence to substantiate murder prosecutions is seldom obtained as the
opposing tribes do not know and therefore cannot identify each other. However,
numerous convictions are obtained in the chief's courts for cattle raiding and
sentences range from six months to four years in prison. Both the Karamojong
and the Suk normally serve their sentences in the Protectorate prison at Moroto.
They work well, seldom give trouble and, compared with other tribes, are model
prisoners. Many of them who have served fairly long sentences and have become
accustomed to wearing the drab uniform of the prison, ask when due for release
to be allowed to retain their uniforms to wear at home. The answer to this
extraordinary request has always been the same-"Naked you came and naked
you go". A very sensible attitude, I submit, as these people are still compara-
tively unsophisticated and no one with any understanding and sympathy would
wish to see them ashamed of their nakedness.
When I say 'understanding and sympathy' I mean just that, as no one who
has spent any length of time in Karamoja could fail to like these simple tribes-
men and sympathize with them-a people whose only love is cattle and who
are entirely honest and straightforward in their own way (except, of course,
when dealing with stolen cattle!) It is a fine thing to have them look one straight
in the face and not only say what they think but mean it. They fear no one and
their absolute trust and confidence in the words of the Government Officer that
they will get a square deal is a source of great pleasure to those of us who have
served in this fascinating district, one of the last remaining pieces of old Africa.
But what of this traditional bitterness between the two tribes? Is it to go on
or is there a solution? Would it be feasible to move the Suk back to Kenya
whence they came, to the other side of the Karasuk and Charangani Hills? They
have established a claim of right now where they are and would fight bitterly
against going back to the even worse erosion and strangeness of Kenya where
they would not find grazing in sufficient quantities. Turn them into agricul-
turalists and make them give up their only interest in life-their cattle? The
Suk say they would die rather than do this, and, truly, it is next to impossible to
get anything to grow in the south-eastern part of Karamoja. They cannot be
moved and bloodshed results if they stay; what, then, is the solution, if there
is one?
I had most interesting conversations with a chief named Lorika on this
subject. A Toposian (Sudanese) by tribe, he was an outstanding leader with
astonishing breadth of vision and depth of intelligence, considering his very
humble background. His murder by the Suk in the 1953 disturbances created
a gap that will be felt for many a long year-even the Suk regretted what they
had done in the heat of battle; for he held the unique reputation of having been
the only chief in the history of Karamoja who was loved and respected with
equal sincerity by both tribes. Had he been spared he might well have risen

to great heights and proved himself a man who could not only have wielded the
baton of authority with courage and unswerving conviction for his people, the
Karamojong, but, vastly more important, for the Suk at the same time. His
death was indeed a great loss.
Lorika was convinced that the only solution to the bitter enmity between the
Karamojong and Suk was assimilation. This sounds very far-fetched at first, but
Lorika pointed out that in two or three very small parts of the tribal border
areas some of the Suk and Karamojong not only lived in close proximity to each
other but a few had actually intermarried and made a success of their lives. He
contended that this state of affairs had gone on for some considerable time and
as neither community had expressed any strong objections against it, an intelli-
gent, patient and gradual approach to the problem on that basis would be well
worth considering.
Lorika was of a very retiring disposition; however, he was drawn out one
night when I camped with him on a safari connected with cattle-raiding, and
gave his opinion on the most practical approach to assimilation. It was that,
firstly, the basic essentials-the fundamental common interest in cattle-already
existed and that if the process of assimilation were founded on this firm core,
eventual success could emerge.
By eventual success he meant that it should be tackled on a long-term basis
and that there should be no question of attempting speedy results, which, he felt,
would do more harm than good. He thought the first move should be in respect
of those few families who had already assimilated successfully. It should be
made worth their while to form a joint settlement, preferably near a central
tribal border where grazing would be assured. Having done this, local acceptance
of the settlement by both tribes would follow. Participation in social life would
soon be followed by the settlement playing an ever-increasing part in the public
affairs of the community. Social intercourse would naturally result in friend-
ships and marriages, public activities would result in the community playing an
increasingly deeper part in general affairs. The community would have its own
chiefs and elders who would exercise their authority under the direct aegis of
the Protectorate Government and not through the senior chiefs of the other
Various attractive conditions could be made with regard, for instance, to
marriage, one of which could be payment by Government of a proportion of a
bride-price if one partner was Suk and the other Karamojong; a reduction in
the bride-price could also be brought about.
Lorika contended that given time, intelligence and patient sympathy a Suk-
Karamojong communal settlement would increase beyond recognition in less
than twenty years and that in ten or twenty more the problem could be solved
permanently. He said the tribe should be called the Karasuk a very apt name
and one which has been the name of a range of hills in south-east Karamoja
since the days before the Suk settled there.
Would this new Karasuk tribe continue the cattle raiding of its forefathers?
Well, with good and firm chiefs much could be done to set a young community
on the road to enlightenment and to reduce bride-price or at least to confine
petty stock-thieving to minor proportions, whilst steps were being taken to solve

the problem of spear-blooding. As there would no longer be a common enemy
much of the incentive to large-scale raiding would cease to exist.
Lorika was an idealist and perhaps his idealism led him to underestimate
many of the technical difficulties which assimilation would present. For instance,
Government assistance to the extent of paying part of the bride-price in cattle
is very optimistic, but like many other simple people Lorika thought Govern-
ment's purse bottomless. The idea is good in its foundation, but the technical
difficulties are impressive and may well be insurmountable..
However that may be, Karamoja remains one of the few parts of the 'old'
Africa of yesterday, the Africa of the camp-fire and the hunt of big game, the
land of the rolling plains and grazing herds, where life stages an exciting and
carefree sequence of cattle-raiding youth, contented and tolerant middle age and
benign indulgent old age, in which the old men love to sit around the fire and
recall the raids and battles of their youth, how they blooded their spears and
acquired their first cattle. Gazing at passing herds, which doubtless contain a
number of stolen cattle, they compare them with the cattle of their youth-how
much better they were and how much more worthwhile 'acquiring'.



N his note on the proposals for a Federal Capital for Eastern Africa which
appeared in Volume 2 of The Uganda Journal, Mr. H. B. Thomas referred
briefly to the transfer on 1 April 1902, of the former Eastern Province of Uganda
to the East Africa Protectorate.' The area involved lay approximately between
the present eastern frontier of Uganda and a line running northward from a
point where the boundary between British and German territory crossed longi-
tude 36 east and turning north-east to include Lakes Naivasha, Elmenteita,
Nakuru and Baringo in the Uganda Protectorate before veering gradually
north again to follow the line of longitude 360 45' east. At the time when Mr.
Thomas wrote the reasons for the transfer had not been made public and there
is still, twenty years later, considerable uncertainty about them. The opening of
the Foreign Office records to 1902, however, has now provided the answer and
has largely confirmed Mr. Thomas's surmises. The interesting feature of the
whole affair is the fact that the Foreign Secretary, the Marquess of Lansdowne,
preferred to act upon the advice of a Foreign Office official, Sir Clement Hill,
instead of accepting the recommendations of so experienced a man as Sir Harry
Johnston, Special Commissioner in Uganda, and in spite of the views expressed
by Sir Charles Eliot, Commissioner of the East Africa Protectorate.
Serious concern had been felt in England at the news of the mutiny of some
of the Uganda Protectorate's Sudanese troops in 1897-8 and a critical article in
the Weekly Times of 27 January 1899, had produced a far from convincing
retort from the pen of Colonel Trevor Ternan in Uganda.2 The appointment of
Sir Harry Johnston as Special Commissioner to reorganize the administration
of the Protectorate can scarcely, therefore, have come as a surprise. In addition,
the near completion of the railway from Mombasa to Lake Victoria, and the
greater ease of communication between Uganda and Europe which this would
entail, suggested that this was an opportune moment for reconsidering the
situation in East Africa though, needless to say, the importance of economy
was duly stressed.3
The part of Johnston's instructions which particularly appealed to his
imagination was that which enjoined him to take special care in choosing the
site for an administrative capital in view of the possibility that the East Africa
and Uganda Protectorates might be merged. As early as February 1900, he had
had a vision of a new empire in East Africa and was writing enthusiastically
to the Marquess of Salisbury about a site he had discovered on the Mau Plateau

1 H. B. Thomas. Uganda J., 2 (1935), pp. 247-9.
2 F.O.2./201. Col. T. Ternan to the Editor of The Times, Kampala, 27 March 1899.
3 F.O.2./200. Foreign Office to Sir Harry Johnston, 1 July 1899.

near mile 475 on the Uganda Railway.4 A surprisingly cool reply to his despatch,
for, after all, Johnston had only been carrying out the terms of his commission,
might have warned him that someone at the Foreign Office was less enthusiastic
about federation that he was himself.5 But what he saw during the remainder of
his stay in East Africa only served to strengthen his views. When, therefore, he
submitted his general report on Uganda to the new Foreign Secretary, the
Marquess of Lansdowne, on 10 July 1901,6 he strongly urged the theme of
Closer Union between the two Protectorates. Remarking upon their inter-
dependence and the close similarity of their interests he said, nevertheless, that
he did not propose their immediate and absolute fusion. Nor, for reasons con-
nected with the distribution of the various tribes, did he recommend any change
either in the direction of extending the rule of Mombasa west of its existing
limits or the rule of Entebbe over any part of the East Africa Protectorate.
Instead of changing provincial or district boundaries he recommended that the
two Commissioners ordinarily residing at Entebbe and Mombasa should con-
tinue with their existing functions, but, in addition, there should be a High
Commissioner who should have supreme control over the policy and finances
of the two Protectorates or of the fused Protectorate. The High Commissioner
should have his headquarters at the new capital which Johnston had roughly
laid out on a site near mile 497 on the Uganda railway and which he proposed
to call King Edward's Town. The site he had earlier recommended, near mile
475, had proved less suitable than he had at first thought. Allowing his enthusi-
asm full rein, Johnston then continued, "This place would be the Simla of
British East Africa, the headquarters of a rule which would stretch from
Kismayu and Mombasa to Gondokoro and the Semliki. Here, and not at
Zanzibar, which knows nothing of the affairs of Inner Africa, should be estab-
lished the Supreme Court. Here should be situated the Treasury and the head
office of Accounts for the two Protectorates. Here, also, should be established
the headquarters of the armed forces and the Medical Department. The High
Commissioner would sometimes visit his Calcutta at Entebbe and his Bombay
at Mombasa. He could refresh himself with ice and snow on Kenia, on Elgon
and on Ruwenzori but his chief home would be at King Edward's Town, in
the heart of a real European Colony on the Mau Plateau within sight of the
Victoria Nyanza and within thirty-six hours' journey of Mombasa.7 To round
off the scheme and to concentrate the attention of the Administration upon the
mainland Protectorates Johnston concluded by recommending that Zanzibar
and Pemba should be severed from British East Africa. For political reasons,
however, he allowed that it might be advisable to give the High Commissioner
some controlling power over the policy followed by the Zanzibar Sultanate.
That Johnston's scheme never bore fruit was mainly due to the intervention
of Sir Clement Hill, Superintendent of the African Protectorates administered
by the Foreign Office. Towards the end of 1900 Hill visited East Africa to
investigate personally the conduct of administration there. Like Johnston he
4 F.O.2./297. Sir Harry Johnston to the Marquess of Salisbury, 18 February 1900.
5 F.O.2./295. Foreign Office to Sir Harry Johnston, 25 April 1900.
6 F.O.2./462. General Report by Sir Harry Johnston on Uganda, 10 July 1901.
7 F.O.2./462. General Report by Sir Harry Johnston on Uganda, 10 July 1901.

was not uniformly impressed by all he saw, particularly in the East Africa Pro-
tectorate. But the solution which Hill proposed differed radically from the
.grand schemes envisaged by Johnston. On his return to England he submitted
a memorandum to the Marquess of Lansdowne on 14 May 1901, setting out the
impressions he had formed while in East Africa. On 25 July he followed this up
with proposals to meet the situation. By this time he had read Johnston's report
but had rejected its conclusion. He agreed with the Special Commissioner that
the weak spot in the existing system was the fact that the Commissioner of the
East Africa Protectorate was also His Majesty's diplomatic representative in
Zanzibar. Unlike Johnston, however, Hill assumed that for the time being at
least that situation must remain unchanged. The mainland administration must,
therefore, be strengthened and the Commissioner, together with his more
important officials, must be free to move about and supervise the activities of
the whole Protectorate without becoming too deeply involved in routine adminis-
tration. Federation as a possible answer to this problem Hill immediately ruled
out on the ground that Uganda's communications were not yet adequate to
enable one man to supervise so large an area effectively. Furthermore, Uganda
in Hill's opinion would probably look increasingly towards the Sudan. To link
the two East African Protectorates at that stage might, in the circumstances,
force trade along the eastern route when its natural tendency might prove to be
to follow the course of the Nile. Possibly, when means of accommodating the
headquarters of the various departments at a new capital in the highlands
seemed more feasible, the question of moving them might be considered. But
for the time being Hill was firmly of the opinion that the East Africa Protec-
torate's headquarters should remain at Mombasa.
Hill next went on to take a more positive line by proposing to adjust the
boundaries between the two Protectorates. The new boundary should start at a
point on the shore of Lake Victoria a little north of Kisumu and should run
north-east along the crest of Mount Elgon until it joined the Turkwell River
which it would then follow northward. To deal with the administration of the
ceded Province the new post of Deputy Commissioner should be created in the
East Africa Protectorate and Hill added the recommendation that the appoint-
ment should be offered to Mr. F. J. Jackson, Acting Commissioner in Uganda,
on account of his intimate knowledge of the Nandi and Masai peoples. The
great advantage of the new scheme, in Hill's opinion, apart from its avoidance
of federation, was that it would make possible the placing of the whole territory
traversed by the Uganda Railway under one Administration. In addition, the
presence of the Deputy Commissioner in the new Province would ensure closer
supervision than had been found possible in the past by the Uganda Administra-
tion which had been primarily concerned with affairs further west.8
It is some indication of the influence exerted by Hill's views that all reference
to the appointment of a High Commissioner and the federation of the two
Protectorates was omitted from the printed version of Johnston's report sub-
mitted to the two Houses of Parliament.9 This did not prevent that great radical
8 F.O.2./519. Memorandum by Sir Clement Hill, 25 July 1901.
9 Command Paper, Africa. No. 7 (1901). I am grateful to Mr. H. B. Thomas for
drawing my attention to the alteration in the printed version of Johnston's report.

and imperialist, Sir Charles Dilke, from speaking in support of a High Com-
missioner in the House of Commons on 19 July, and Johnston himself was far
from being convinced that Hill's proposals were better than his own. On
9 August 1901, therefore, he submitted to the Foreign Secretary a more detailed
statement of his views. In it he expressed his complete disagreement with the
idea of transferring any part of the Uganda Protectorate to its neighbour. He
recognized that Uganda was still centred upon the kingdom of Buganda while
the affairs of the East Africa Protectorate radiated from the Arab coast. But he
insisted that the major part of Uganda must continue to look to the coast rather
than to the Sudan for its link with the outside world. In addition, he believed
that the transfer of Uganda's Eastern Province would involve the severance of
the tribes of that area from their natural focus in Uganda. This territory, he
maintained, could be more easily administered from Uganda and, as to the
question of placing the whole railway under one administration, surely that was
admirably covered by the idea of a central government at Mau. From the point
of view of efficient administration, moreover, it was clear that to retain the
existing officers was a far better policy than to introduce men from the East
Africa Protectorate's administration whose outlook would inevitably be
governed by their connection with the coast and who would take time to under-
stand the problems of the new Province.10
Johnston's letter, however, reached the Marquess of Lansdowne too late to
play a decisive part in influencing the Foreign Secretary's conclusions. For, on
10 August 1901, discussions took place in Lansdowne's office which resulted
in the acceptance of Hill's proposal. Hill's comment, when he did see Johnston's
letter, indicated a complete lack of sympathy with Johnston's views. There was
little in the letter which called for remark, he wrote, and he believed all
Johnston's arguments had already been answered. Somewhat ungraciously he
added that if the Marquess of Lansdowne wished to reconsider the whole issue
he was, nevertheless, prepared to go into each point separately once again."
Viscount Cranborne, Under-Secretary of State for Foreign Affairs, was unwilling
to dismiss Johnston's opinions quite so summarily. The Special Commissioner,
he felt, held strong views which were worthy of consideration, and if federation
was contemplated in two or three years' time any intermediate step would be
futile. He recognized the force of Sir Clement Hill's arguments, however, and
appreciated that Hill's plan would involve a minimum of disturbance.12
The Marquess of Lansdowne, it seems clear, regarded the decision as having
already been taken. But he felt he would like the confirmation of Sir Charles
Eliot who already, in June, had expressed the opinion that the East Africa
Protectorate, even with its existing boundaries, was too big for effective adminis-
tration in view of the purely local character of the indigenous political systems."
On 26 August 1901, the Foreign Secretary wrote to Sir Charles explaining that
he was in favour of transferring the territory east of Mount Elgon to the East
1o F.O.2./463. Sir H. H. Johnston to the Marquess of Lansdowne, Saumarez Park,
Guernsey, 9 August 1901.
11 F.O.2./463. Minute by Sir Clement Hill, 17 August 1901.
12 F.O.2./463. Minute by Viscount Cranborne. Undated.
13 F.O.2./448. Report on the East Africa Protectorate by Sir Charles Eliot, 10 June 1901.

Africa Protectorate with a view to improving the administration of the British
possessions in East Africa and asked the Commissioner to give his opinion.14 In
reply, Eliot stated at once that he, along with the majority of the officials in
East Africa, was impressed by Johnston's proposals for merging the two
Protectorates. While accepting Hill's argument that poor communications would
render impossible the supervision of so large an area by one man for the time
being, he hoped the idea of federation would not be wholly forgotten. As a
temporary alternative he thought the measure proposed was eminently desir-
able, particularly as the existing boundary cut the lands of the Masai in two. In
addition, it would be clearly advantageous to place the whole of the railway
under one administration. He could not, however, accept Sir Clement Hill's
suggestion that no effort should be made to induce the trade of Uganda to
follow the Mombasa route. It seemed clear to him that unless very energetic
efforts were made in that direction the railway would soon be in a disastrous
financial condition.15
While awaiting Eliot's reply, the Foreign Secretary held a number of con-
versations with Sir Harry Johnston but he was already convinced by Hill's
arguments. Johnston was conscious that he was making little headway and
wrote to Viscount Cranborne on 6 November 1901, in the hope of driving
home the views which he felt he had been unable to make sufficiently clear in
the course of his verbal exchanges with Lansdowne. Once again he insisted that
he could see nothing to be gained from a transfer of territory between the two
Protectorates. For, he said, if the Commissioner continued to act as Agent and
Consul-General in Zanzibar the addition of further territory to the East Africa
Protectorate would present him with a task too great for one man to contem-
plate. If, on the other hand, Zanzibar was removed from his sphere of responsi-
bility and the Commissioner took up residence in the interior it would amount
to accepting the idea of a High Commissioner without developing it to its logical
conclusion. What Johnston wished to make clear was the difference between
the task of supervising a large area which would fall to the High Commissioner
and which would be quite feasible and, on the other hand, expecting one Com-
missioner to carry out the actual business of administering the enlarged Pro-
tectorate suggested by Hill. He seems to have been unaware of Hill's proposal
that the new Province should be administered by a Deputy Commissioner.16
Johnston's fear that his arguments were making little impression on Lans-
downe were confirmed by the Foreign Secretary's minute of 8 November, 1901,
in response to the Special Commissioner's letter. "I don't think there is really
anything in this to stand in the way of the transfer of the Eastern Province,"
Lansdowne wrote, "but let us see what Sir Charles Eliot says in the reply to my
telegram of yesterday." In that telegram the Foreign Secretary had emphasised
that the idea of amalgamating the two Protectorates had not been ruled out
permanently and that he did not wish to take any step which might jeopardize
14 F.O.2./443. Marquess of Lansdowne to Sir Charles Eliot, 26 August 1901.
15 F.O.2./450. Sir Charles Eliot to the Marquess of Lansdowne, Zanzibar, 10 October
16 F.O.2./464. Sir H. H. Johnston to Viscount Cranborne, the Mount, Shere, Surrey,
6 November 1901.

the ultimate prosecution of the larger scheme. He was anxious therefore to
obtain Eliot's opinion as to whether Hill's proposal could be regarded as an
instalment of the larger project.17 In reply, Eliot welcomed the scheme for the
transfer of territory as far as it went, mainly on the ground that it would force
the East Africa Protectorate to take a greater interest in the interior and in the
affairs of Uganda and thereby pave the way for the unification of the two
Protectorates.'1 So Hill's scheme was adopted,19 though apparently on the
ground that it was only a transitional stage which in due course would lead to
the federation of the two Protectorates so earnestly desired by Johnston and
Eliot. Mr. C. W. Hobley was given the task of defining the new frontier more
accurately and numerous problems arose owing to the fact that Hill's 'natural'
frontier did not readily coincide with tribal boundaries.20 But, as Mr. Thomas
has written, "The rites of the divorce of the old Eastern Province from Uganda.
and the settlements on her remarriage to East Africa were conducted at the end
of November at Njoro."21

17 F.O.2./455. Telegram from Lansdowne to Eliot, 7 November 1901.
18 F.O.2./456. Telegram from Eliot to Lansdowne, 8 November 1901.
19 Estebbe Archives. S.M.P. No. 9074. Lansdowne to Eliot, 5 March 1902.
20 Entebbe Archives. S.M.P. No. 9074. Eliot to Lansdowne, Nairobi, 25 January 1902.
21 H. B. Thomas. Uganda J., 2 (1935), 248. See also H. B. Thomas and A. E. Spencer,
A History of Uganda Land and Surveys. Entebbe, 1938. Pp. 7-9.

Barbus Cuvier and Cloquet 1816
Body somewhat compressed, the scales generally large. Cheeks not covered
by the sub-orbital bones. Lateral line usually situated nearer to the ventral
than to the dorsal body profile, but always running along the middle of the
caudal peduncle. Mouth terminal in position, the lips showing considerable
variation in their degree of development. No flap of skin anterior to the upper
lip. Circum-oral barbels present, or, less frequently, absent. Dorsal fin with the
last unbranched ray often ossified and spine-like. Pharyngeal teeth in three
series, the majority being hooked or spoon-shaped, although one or more
anterior teeth in the inner series may be molar-like.
Undoubtedly more species have been referred to Barbus than to any other
fresh-water, non-cichlid genus in Africa. Intra-specific variability in this genus
is, however, extremely high, with the result that many so-called species should
probably be considered as geographical variants, or, atypical specimens of some
other species. There is an urgent need for a thorough taxonomic revision and
field study of the genus. When this revision has been carried out, some of the
specific names given below will doubtless be changed as the result of several
"species" being amalgamated. Also, it seems likely that a few new species will
be described.
Very little is known about the ecology and feeding habits of the small species
of Barbus (those belonging to the sub-group with radiately striate scales). Most
of the information given here is based on the examination of very few specimens.
The key given below is, in many respects, inadequate. However, it should
enable the commoner species of Uganda Barbus to be identified with a fair
degree of certainty.

This key is based on adult fishes and may prove misleading if used to identify
young individuals. For example, two important 'key-characters', the barbels
and the last unbranched ray of the dorsal fin, change with the size of the fish.
Barbels tend to be relatively shorter in young fishes whilst the ray is but slightly
ossified and weakly, if at all serrated in young fishes.
Lateral line present 1
Lateral line absent, or, represented only by a few anterior scales 2
1 Exposed surface of the scales with numerous longitudinal striae A
Exposed surface of the scales with radiating striations which converge
at a point near the scale centre R

A Last unbranched ray of the dorsal fin strong and bony but shorter
than the head; depth of body not markedly greater than the head
length B. altianalis1
Last unbranched ray of the dorsal fin, strong, bony and as long as,
or, longer than the head; depth of body greater than the head length
B. bynni
B Last unbranched ray of the dorsal fin bony and serrated on its
posterior face (a)
Last unbranched ray bony and enlarged but without serrations on the
posterior face (b)
Last unbranched ray neither bony nor enlarged (c)
(a) Lateral line with 34-38 scales Barbus cf. B. amphigramma
Lateral line with 25-30 scales (ii)
(ii) Anterior barbel extending to a point about mid-way along the
posterior barbel. Base of the pelvic fins below the first dorsal
rays B. portali
Anterior barbel extending only to the base of the posterior
barbel, or, very slightly beyond. Base of the pelvic fins entirely
in advance of the first dorsal ray.
a No silver lateral streak (Lake Victoria and its affluent streams)
B. minchini
/3 Silver lateral streak (Streams of western Uganda, Lakes Edward,
Bunyoni and Mutanda) B. kersteni
(b) Dorsal fin with 8 branched rays; lateral line with 36-39 scales
B. nummifer
Dorsal fin with 10 branched rays; lateral line with 33-34 scales
B. somereni
(c) Barbels well developed (i)
Barbels small, anterior barbel sometimes absent (ii)
(i) Three dark spots on the flanks B. prince
A broad, silver band running above the lateral line B. neglectus
(ii) Both anterior and posterior barbels minute; lateral line almost
straight; 3 scales between the lateral line and the origin of
dorsal fin B. doggetti
Anterior barbel absent, posterior barbel from minute to j of
the length of the eye; lateral line curved downwards on the
flanks; 5 or 6 scales between the lateral line and the origin
of the dorsal fin B. magdalenae

2 Lateral line absent; a single barbel on each side of the mouth; last
unbranched dorsal fin ray bony and weakly, or, strongly, serrated
on its posterior face B. apleurogramma
1 See under B. altianalis in the text for notes on three Lake Kyoga species, B. obesus;
B. kiogae; and B. longirostris.

Barbus altianalis Boulenger 1900. (Fig. 29)
English: Barbel (particularly Ripon Falls Barbel) or Barbus. Native names:
Kisinja, Nkuyu (Lake Victoria); Ngambwa (Lunyoro, Luruli, Lukenyi, and
Lunyara); Sanga (Ludope); Changa (Lango).

FIG. 29
Barbus altianalis: adult. (Drawn by Barbara Williams.)
Description: Depth of body equal to, or slightly greater than the length of
the head, which is contained 3-4J times in the standard length. In some fishes
the head does not pass imperceptibly into the body; instead, there is a well-
marked hump at the nape. Two pairs of barbels present, the anterior somewhat
shorter than the eye, the posterior as long as the eye. Snout rounded or some-
what acute, its length contained 21-3 times in the head length. Mouth inferior;
lips varying in shape and size from thin to fleshy, the latter type often produced
into large upper and lower lobes. Fishes with well-developed lips often have the
snout rather fleshy. This variation in lip size is not a sexual difference, but it
does seem likely that thickened lips are more pronounced in older and larger
fishes. Eye diameter contained 3 (in young) to 7 times in head length. Dorsal
fin with III-IV, 8-9 rays, the last unbranched ray strong and bony, from I to w the
length of the head. The base of the pelvic fins is situated below the anterior
rays of the dorsal fin. Lateral line with 32-39 scales; 51-61 scales between the
lateral line and the first dorsal ray; 21-31 scales between the lateral line and the
pelvic fin base.
Young B. altianalis do not closely resemble the adults (Fig. 30).
Three sub-species of B. altianalis are recognized (Worthington, 1932a),
namely: B. a. altianalis Blgr. (Lake Kivu and the Ruzizi River), B. a. radcliffi
Blgr. (Lakes Victoria and Kyoga, and the Victoria Nile) and B. a. eduardianus
Blgr. (Lakes Edward and George). The characters separating these sub-species
are slight, and for the purpose of this paper it is sufficient to use geographical
distributions as diagnostic characters.
Coloration: Tarnished silver in young fishes, becoming golden-green in
Size: The largest recorded specimen caught by angling was 90 cm. long and
weighed 19 lb. (E.A.F.R.O.). Fishes weighing over 40 lb. have been caught in
gill-nets and on long-lines (Uganda Government, 1948).


2 2c.m.
FIG. 30
Barbus altianalis: juvenile. (Drawn by Barbara Williams.)
Habitat: Lake Victoria: Inshore waters and in the affluent rivers (E.A.F.R.O.
and Graham, 1929). The species is particularly abundant in the Victoria Nile.
Before the Ripon Falls were submerged, large numbers of B. a. radcliffi congre-
gated in the turbulent waters at the foot of the falls. Many individuals could
be seen leaping and attempting to leap the falls. Nowadays a similar concentra-
tion of fishes is found at the Owen Falls Dam. It is not known whether there
was ever an important movement of fishes from the river to the lake, although
this has been suggested.
An interesting idea, held by some African fishermen, is that the Barbus
ascended the falls and moved to the Kagera river (Tanganyika-Uganda bound-
ary) to spawn. The fishes were believed to find their way by following the stream
of water which flowed into the lake from the river and maintained its indi-
viduality until it passed over the Ripon Falls.
Lake Kyoga: The more open waters of the lake. (Worthington 1929a.)
Lakes Edward and George: Shallow water. Common in Lake George and
the Kazinga Channel. (Worthington, op. cit.)
Food: Varied, including Mollusca, aquatic vegetation, fishes (particularly
small Haplochromis) and insect larvae (Graham, op. cit., and E.A.F.R.O. for
B. a. radcliffi, and Worthington, op. cit., for B. a. eduardianus).
Breeding: There is no positive evidence on the breeding sites or seasons of
B. altianalis. Indirect evidence suggests that, at least in Lake Victoria, the
species breeds in rivers and large streams during flood periods.
Sport: Barbus altianalis radcliffi, the so-called Ripon Falls Barbel, provides
excellent sport. An abundance of small bones detracts from its palatability.
Economic importance: Although contributing to the catches of commercial
fishermen, the species is not of great economic importance.
Distribution: Uganda: Lakes Edward, George, Victoria and Kyoga. Else-
where: Lake Kivu, the Ruzizi river.
Barbus obesus Worthington 1929; B. kiogae Worthington 1929, and
B. longirostris Worthington 1929 (Fig. 31).
These three species were described by Worthington (1929b) from specimens
caught in Lake Kyoga, where the species are apparently endemic.

Compared with typical Barbus altianalis radcliffi, Barbus obesus has a
deeper body (2 times in body length) and consequently differs in outline. Barbus
kiogae and B. longirostris differ, respectively, in having a marked hump at the
nape and a slightly longer snout. In both species the mouth is terminal.
The validity of these species can only be checked when more specimens,
together with data on their ecology, are available from Lake Kyoga. At present
it can be said that specimens of B. altianalis showing many of the prominent
characters of B. kiogae and B. longirostris have been caught in Lake Victoria.
Barbus obesus is known from a single specimen and it is possible that this
species, together with B. kiogae, and B. longirostris may only represent more
extreme variants of the highly variable species, B. altianalis.

FIG. 31
a Barbus obesus; b B. kiogae; c B. longirostris; outlines from Worthington (1929b).

Barbus bynni (Forsk.) 1798. (Fig. 32)
English: Barbel. Native names: Kisinja (Lunyoro and Lugungu); Sirri (Alur);
Oshoi (Jonam).
Description: Similar to Barbus altianalis except that the body is deeper (its
depth contained 21-31 times in the standard length) and the last unbranched
dorsal ray is stouter and from 1-1 longer than the head; in a few exceptional
specimens the spine was found to be slightly shorter than the head.

FIG. 32
Barbus bynni. (Drawn by Barbara Williams.)

Coloration: Silvery to yellowish. Anal and caudal fins tinged with pink, all
other fins yellow-orange.
Size: The largest recorded fish was 82 cm. long and weighed 12 pounds
(Worthington, 1929a).
Habitat: Inshore waters of Lake Albert; also in the Murchison Nile. Worth-
ington remarks that B. bynni was not found in large numbers in Lake Albert,
and that it was apparently more abundant in rivers.
Food: Aquatic plants, mollusca and insects (Worthington, op. cit.).
Breeding: Unknown.
Sport: Hurcomb (Angling in the Sudan; n.d.) does not write enthusiastically
on the sporting qualities of this species.
Economic importance: B. bynni contributes to the catches of native fisher-
men, but, like B. altianalis, it cannot be rated a species of great commercial
Distribution: Uganda: Lake Albert, the Murchison and Albert Niles. Else-
where: Nile and Lake Rudolf. A distinct sub-species, B. b. rudolfianus has been
described from the latter locality (Worthington, 1936).

Barbus cf. B. amphigramma Boulenger 1903. (Fig. 33)
Native name: Included in 'Kisinja' the general term for all small Barbus.
The identity of this species must remain doubtful until more is known about
the Barbus of Africa. It resembles both Barbus amphigramma and a species,
B. macropristis Blgr. 1904, which Boulenger later (1911) considered to be

identical with the widely distributed B. paludinosus Peters 1852. Furthermore,
the species B. thikensis, which has been included in faunal lists of Lake Victoria
(Cunnington, 1920; Graham, 1929), is probably identical with B. amphigramma
(Dr. Trewavas, in litt.).
Thus, at least three species previously recorded from Lake Victoria (and
Uganda) are probably represented by the single species here referred to as
B. amphigramma.
Description: Depth of body contained 31-4 times in the standard length,
length of head 4-4 times. Snout fairly acute, or somewhat rounded, its length
equal to, or slightly longer than, the eye, which is contained 31-4 times in the
head length. Anterior barbels 1-5 the length of eye, posterior barbels J-:. Dorsal
fin with III, 7 rays, the last unbranched ray bony, strongly serrated on its
posterior face and as long as, or, longer than the head. Pelvic fin base entirely,
or partly, in advance of the first unbranched dorsal ray; tip of the pectoral fin
reaching the base of the pelvic fin. Lateral line with 34 (rarely)-38 scales;
6-71 scales between the lateral line and the first dorsal fin ray; 3-4 scales
between the lateral line and the pelvic fins.
Size: Rarely exceeding 130 mm. standard length.

FIG. 33
Barbus amphigramma. (Drawn by Barbara Williams.)
Coloration: Greyish-green above, silver below; a dark mid-lateral band and a
thin, interrupted black line on the anterior third of the lateral line.
Food: Insect larvae (E.A.F.R.O.).
Habitat: B. amphigramma is frequently found in small streams and rivers
flowing into Lake Victoria. The species also occurs in the inshore areas of the
Distribution: Uganda: Lake Victoria and its affluent rivers and streams; Lake
Kyoga and streams flowing into the lake. Elsewhere: Barbus amphigramma
(sensu strict) is recorded from Kenya.

Barbus portali Boulenger 1906. (Fig. 34)
Native name: Unknown, probably included in the general term 'Kisinja'.
Description: A small species, rarely exceeding 110 mm. standard length.
Depth of body contained 31-31 times in the standard length, length of head
3J-4 times. Snout somewhat rounded, its length equal to the diameter of the eye

and contained 3-4 times in the head length. Mouth terminal. Anterior barbels
equal to the diameter of the eye in small fishes and 11-12 times longer than the
eye in larger individuals, extending to about mid-way along the posterior barbel;
posterior barbels 1 to 21 times longer than the eye. Dorsal fin with III, 7 rays,
the last unbranched ray bony and strongly serrated on its posterior face. Pelvic
fin base entirely below the first dorsal rays. Lateral line with 26-31 scales; 41-5
scales between the lateral line and the first dorsal ray; 3) or 4 scales between the
pelvic fin base and the lateral line. It is interesting to note that B. portali from
the Waisoke river system have fewer scales in the lateral line (26-28) than have
specimens from streams flowing into Lake Victoria (30-31 scales).
The three smallest specimens examined (27-31 mm. standard length) were
recognizable as B. portali on adult characters. However, the serrations on the
dorsal fin spine are relatively weak, and, in one fish, the anterior barbel extended
only to the origin of the posterior barbel and not to about mid-way along it.

FIG. 34
Barbus portali. (Drawn by Barbara Williams.)

Coloration: Brownish olive-green above, silver below. Fins faint brownish-
green except the caudal which is pinkish-brown. In young fishes there are three
conspicuous spots or blotches on the flanks. These spots are less intense in
larger individuals, and may even be absent.
Food: Principally insect larvae and aquatic plants (E.A.F.R.O.).
Habitat: B. portali occurs in both permanent, fast flowing streams and in
seasonal streams running into Lake Victoria. Its sudden appearance in these
temporary streams indicates that the species also occurs in Lake Victoria, where
its habitat is unknown.
Distribution: Lake Victoria and affluent streams; Mpanga River; Waisoke
River and tributaries (flowing into Lake Albert); Kirima and Mongiro rivers
(Bwamba district, flowing into the Semliki river); Hima and Malawa rivers.
(Boulenger, 1911 and 1915; E.A.F.R.O., collections made by P. Hainsworth;
Percy and Ridley, 1955.)
Several other species closely resembling B. portali occur in East Africa and
it is probable that further study will show that these 'species', together with
B. portali, are members of one, wide-spread species. A species, B. carpio
Pfeffer, 1896, described from a single specimen caught in Lake Albert is closely

related to, if not identical with B. portali. If further study shows that they are
identical then the species here referred to as P. portali will have to be renamed
B. carpio since this fish was described before B. portali.

Barbus minchini Boulenger 1906. (Fig. 35)
Native name: Unknown, probably included in the general term 'Kisinja'.
Description: A small species not known to exceed 90 mm. in standard length.
Depth of body contained 31-31 times in the standard length, length of head
31-31 times. Snout somewhat rounded, a little shorter than the eye, which is
contained 31-4 times in the head length. Mouth terminal; the anterior barbels
from 1-1 of the eye diameter and reaching the origin of the posterior barbel,
which is as long as, or slightly longer than, the eye. Dorsal fin with III, 7 rays,
the last unbranched ray bony and serrated on its posterior face. Pelvic fin base
entirely in advance of the first dorsal rays. The pectoral fins of male fishes are
more expansive and larger than those of females. Lateral line with 23-26 scales;
44-5 scales between the pelvic fin base and lateral line.

FIG. 35
Barbus minchini. (Drawn by Barbara Williams.)
Coloration: Golden-silver above, shading to silver below; scales above the
lateral line with dark centres. A bright orange spot on the operculum. Caudal
and anal fins bright orange, dorsal fin faint orange; all other fins colourless.
Food: Predominantly insect larvae (E.A.F.R.O.).
Habitat: Streams, both fast flowing and sluggish. Since the species occurs in
temporary streams connected with Lake Victoria, it is assumed that it also
occurs in the lake.
Distribution: Lake Victoria and its affluent streams; the Kyoga system;
Malawa River (E.A.F.R.O. collections; Boulenger, 1915).
Barbus kersteni Peters 1868
Native name: Unknown, but probably included in the general term 'Kisinja'.
Description: Similar to Barbus minchini, differing mainly in its somewhat
smaller eye, more rounded snout, and the presence of a bright silver mid-lateral
stripe, which may be interrupted.
Coloration: Silvery-bronze above, silver below; a very distinct, silver mid-
lateral stripe, which darkens after death but is still clearly silver.

Distribution: Uganda: Kazinga Channel, Lake Edward; Semliki River; Lakes
Bunyoni and Mutanda (Uganda Game and Fisheries Department collection);
probably also in Lake Kyoga (Worthington, 1929a). Elsewhere: Mount Kili-
manjaro; Umbugwe, Masai district, Tanganyika Territory; Sagana, Kiwale
and Kibwezi rivers, Kenya (Kenya Game Department collections). The record
of this species from Komati Poort, Transvaal (Gilchrist and Thompson, 1913)
is open to some doubt.
A very similar species, Barbus luhondo, which Boulenger (1916) considers
"doubtfully distinct from B. kersteni" is known from Lake Luhondo, in north-
western Ruanda-Urundi.

Barbus nummifer Boulenger 1904. (Fig. 36)
Native name: Unknown.
Description: A small species, not exceeding 135 mm. standard length.
Depth of body contained 31-4 times in the standard length, length of head
4-4z times. Snout somewhat rounded, as long as, or slightly longer than, the eye,
which is contained 3J-41 times in the head length. Mouth terminal or slightly
inferior, the anterior barbels as long as the eye; posterior barbels from 1i-1l
times longer than the eye. Dorsal fin with III, 8 rays, the last unbranched ray
very strong, bony but not serrated on its posterior face. Base of the pelvic fins
below, or partly in advance of, the first dorsal rays. Lateral line with 36-39
scales; 5-61 scales between this line and the first dorsal ray; 4 scales between
the pelvic fins and the lateral line.

FIG. 36
Barbus nummifer. (Drawn by Barbara Williams.)

Coloration: Sandy above, silvery below. Three well-defined black spots on
the flanks, the first spot slightly in advance of the dorsal fin, the second some-
what behind this fin and the third near the origin of the caudal fin.
Habitat: Streams, both temporary and permanent; also in Lake Victoria,
where its habitat is unknown.

Food: Insect larvae and bottom debris.
Distribution: Uganda: Lake Victoria and its affluent streams; Malawa river.
Elsewhere: Bubu river (flowing into Lake Manyara) Tanganyika Territory.

Barbus somereni Boulenger 1911
Native name: Unknown.
Description: Depth of body contained 32-4 times in the standard length,
length of head 4-4 times. Snout rounded, 13 times longer than the eye, which
is contained 5 times in the head length. Mouth terminal, the anterior barbel as
long as the eye, the posterior barbel from 11-11 times longer than the eye.
Dorsal fin with III, 10 rays, the last unbranched ray, bony and enlarged but not
serrated on its posterior face. Pelvic fin base entirely behind the first dorsal ray.
Lateral line with 33-34 scales; 51 scales between this line and the first dorsal
ray; 3 scales between the pelvic fin and the lateral line.
Coloration: Sandy above, silver below.
Distribution: This species is known only from a few specimens collected in
the following Ruwenzori streams; Sebwe river, Tokwe river; Kirimia river
(Boulenger, 1916, Percy and Ridley, 1955). It does not apparently ascend above
5,500 ft. B. somereni reaches a larger size (235 mm.) than other Uganda Barbus
belonging to the sub-group with radiately striate scales.

Barbus prince Riippell 1837
Native name: Unknown.
Description: A small species, not known to exceed 110 mm. in length.
Depth of body contained 23-3 times in the standard length, length of head
3-4 times. Snout rounded, shorter than the eye, which is contained 3-31 times
in the head length. Mouth terminal; the anterior barbels from I to as long as the
eye, posterior barbels equal to, or, 1I times greater than the eye diameter. Dorsal
fin with III, 7-8 rays, the last unbranched ray not enlarged and not bony. Pelvic
fin base below the first dorsal ray. Lateral line with 28-32 scales; 42-5) scales
between this line and the first dorsal ray; 2 or 3 scales between the pelvic fins and
the lateral line.
Coloration: Greenish-silver above, silver below. Three distinct black spots
on the flanks; in life these spots are less intense and may even be absent. Fins
pale yellow.
Ecology: Unknown.
Distribution: Uganda: Kazinga Channel, Lake Edward (Worthington,
1932b); Lake Albert (E.A.F.R.O. collection). Elsewhere: Semliki river (Poll
and Damas, 1939); the Nile.

Barbus neglectus Boulenger 1902
Native name: Unknown.
Description: A small species not known to exceed 55 mm. standard length,
and similar to B. prince, except for the presence of a broad, silver lateral stripe.
The occurrence of this species in Uganda is based on a single specimen col-
lected by E.A.F.R.O. in the Albert Nile at Rhino Camp. Otherwise, B. neglec-
tus is known from the lower and Blue Niles.

Barbus doggetti Boulenger 1904
Native name: Unknown, but probably included in the general term 'Kisinja'.
Description: A small species known only from a few specimens caught in
Lake Victoria. It is characterized by having very small barbels, and a slender,
unossified last unbranched dorsal fin ray.
Barbus magdalenae Boulenger 1906. (Fig. 37)
Native name: Unknown, but probably included in the term 'Kisinja'.
Description: A small species, not known to exceed 70 mm. in standard length.
Depth of body contained 31-4 times in standard length, length of head con-
tained 4-5 times. Snout variable, from strongly to moderately rounded, its
length contained 4-4j times in the head length. Eye large, its diameter con-
tained 2j-3 times in the head length. Mouth terminal; no anterior barbel, the
posterior barbel varying in length from minute to f as long as the eye. Dorsal
fin with III, 8 rays, the last unbranched ray neither enlarged nor bony. Pelvic
fin base slightly anterior to the first dorsal ray. Lateral line with 29-32 scales;
curved, its course on the flanks nearer the ventral than the dorsal outline; 5-6
scales between the lateral line and the first dorsal fin ray; 1 or 2 scales between
the lateral line and the pelvic fin base.
The shape of individual B. magdalenae is variable, but the figured specimen
can be considered typical.

FIo. 37
Barbus magdalenae. (Drawn by Barbara Williams.)
Coloration: Tarnished silver, with an intensely silver mid-lateral stripe which
darkens after death and is often black in preserved specimens.
Habitat: Marginal water-lily swamps of Lakes Victoria and Nabugabo, but
running up temporary streams (possibly to breed) during the rainy seasons.
Food: Unknown.
Distribution: Known only from Lakes Victoria and Nabugabo.
Barbus apleurogramma Boulenger 1911. (Fig. 38)
Native name: Included in the general term 'Kisinja'.
Description: A small species rarely exceeding 55 mm. in length.
Depth of body contained 3-3 times in the standard length, length of head

31-3 times. Snout rounded, as long as, or slightly shorter than, the eye which is
contained 3J-3- times in the head length. Mouth terminal; no anterior barbels,
the posterior barbels from -31 the length of eye diameter. Dorsal fin with III,
7-8 rays, the last unbranched ray strong and bony, weakly or strongly serrated
on its posterior face. Pelvic fin base below, or, slightly anterior to the first
dorsal rays. No lateral line; 23-25 scales in a longitudinal series counted along
the line normally occupied by the lateral line scales; 7 scales between the dorsal
and pelvic fins.

FIG. 38
Barbus apleurogramma. (Drawn by Barbara Williams.)
Coloration: Tarnished silver, darkest dorsally; flank scales with dark centres.
A dark spot at the base of the caudal fin. Dorsal and anal fins black, but with a
large, orange blotch in the centre. Caudal fin tinged with orange.
Habitat: This attractive little species is perhaps the most common Barbus
occurring in temporary and permanent streams flowing into Lake Victoria. In
the lake it inhabits the marginal water-lily swamps. It has found its way into
many dams and ponds near Lake Victoria.
Food: Insect larvae (E.A.F.R.O.)
Breeding: Probably in the temporary streams which connect with the lake
during the rainy seasons. (E.A.F.R.O.)
Distribution: Known only from Lake Victoria and associated streams.
Records which require confirmation
Barbus tetraspilus Pfeffer 1896, and B. trispilopleura Boulenger 1902, have
both been recorded from Lake Victoria (see Cunnington, 1920). It seems likely
that B. tetraspilus was included as a result of its locality, which is given as
Kavalli River, Bukoba, Upper Ituri (Congo system), being confused with
Bukoba (Tanganyika Territory) on the shore of Lake Victoria.
The record of B. trispilopleura (Pellegrin, 1910. Mem. Soc. Zool. France, 22,
286) has never been confirmed. The species is otherwise known only from
Abyssinia. In this case, Pellegrin's specimen may have been of a Barbus at
present thought to be an undescribed species.

In addition to the general notes on counts and measurements used in identifica-
tion, and outlined in Chapter I, the following are used when identifying Siluroid
(i) The circum-oral barbels are named as follows: Nasal (situated on the
dorsal surface of the head, near the nostrils); maxillary (originating from
the upper lip); inner and outer mandibulars (situated on the lower jaw).
(ii) The posterior extension of a barbel is determined by laying it back along
the head.
(iii) With the exception of the Clariidae, head length is measured as for the
other fishes. In the Clariidae it is measured along the dorsal surface of
head, from the snout tip to the most posterior point of a triangular pro-
jection from the hind-end of the skull. This process is called the supra-
occipital process.
(iv) In the Mochocidae, reference is made to the humerall process'. This is a
bony, triangular projection from the pectoral girdle. Measurements are
made as shown in Fig. 39.
(v) The vomerine tooth band lies immediately behind the band of pre-
maxillary teeth (upper jaw).

FIG. 39
Sketch showing the measurements of length and depth of the humeral process in
Synodontis. a=depth; b=length.

Family: Bagridae
Body scaleless, moderately elongate. Adipose dorsal fin present. Dorsal and
anal fins short, the former with a well-developed spine. Three or four pairs of
unbranched circum-oral barbels.
Fishes of this family are found in Africa and Asia; two genera occur in

(i) Four pairs of circum-oral barbels Bagrus
(ii) Three pairs of circum-oral barbels (nasal barbels absent) Auchenoglanis

Bagrus Cuvier 1817
Body moderately elongate, slightly compressed. Dorsal and anal fins short,
the former comprising a spine and 8 to 11 branched rays. Adipose dorsal fin
long. Four pairs of circum-oral barbels present.
At least two species occur in Uganda.

Head broad, its width contained 11-11 times in the head length, last ray
of the dorsal fin in advance of the vertical from the last ray of the pelvic
fin B. docmac
Head narrower, its width contained l~-1 times in the head length; last
ray of the dorsal fin behind the last ray of the pelvic fin B. bayad

Bagrus docmac (Forsk.) 1775. (Fig. 40a)
Native names: Semutundu (Lunyoro, Luruli and Lake Victoria); Mboli (Lun-
yara, Lukenyi, Ludope); Obanga (Lango); Omukora (Lunyankole); Oreko
(Lugungu); Ombala (Jonam).
Description: Little formal description of this species need be given (see
generic description, key, and figure) except to note that the barbels show great
variation in length, being proportionately longer in small fishes. Also, the fila-
mentous extension of certain rays in the caudal fin may be absent.

3 c.m.

FIG. 40a
Bagrus docmac. (Drawn by Barbara Williams.)

It seems likely that Bagrus degeni Boulenger 1906, a species known only from
Lake Victoria, is identical with B. docmac. The characters supposedly separat-
ing B. degeni from B. docmac are slight and specimens showing intergradation
between the two 'species' are common (see also Graham, 1929).
Coloration: Dark grey-black above, creamy-white below.
Size: Bagrus docmac attains a large size; specimens measuring over 100 cm.
long and weighing 50 lb. have been recorded from Lakes Victoria and Edward.
Poll and Damas (1939) record seeing the head (more than 40 cm. long) of a
large Bagrus, which they estimate to have weighed 75 kg.
Habitat: Lake Victoria: Widespread in both deep and shallow water (Graham,
op. cit.; E.A.F.R.O.).
Lake Edward: Common in shallow water (Worthington, 1932b). Also found
in the Kazinga Channel and Lake George.
Lake Albert: Inshore and open waters (Worthington, 1929a).
Lake Kyoga: Open waters.
Lake Nabugabo: Ubiquitous (Worthington, op. cit.).
Food: Bagrus docmac is a predator, feeding mainly on small fishes, par-
ticularly Haplochromis. Insect larvae and Crustacea (prawns) are also eaten,
the former especially by young fishes. (Worthington, op. cit. and 1929a; Graham,
op. cit. Poll and Damas, 1939; Hulot, 1956; E.A.F.R.O. unpublished records.)
Breeding: Unknown.
Economic importance: Although not abundant in any Uganda lake, B. docmac
does contribute significantly to the catches of African fishermen in most lakes
except Lake Kyoga (see Uganda Government; and Lake Victoria Fisheries
Service, 1948 et seq.).
Periodic mortality of B. docmac in Lake Victoria: At irregular intervals,
numbers of B. docmac are found floating at the surface, either dead or mori-
bund. Theories put forward to explain this phenomenon are varied and include.
epidemic diseases and underwater eruptions of volcanic gases. At present there
is little or no evidence to support any of these ideas.
Distribution: Uganda: Lakes Victoria, Kyoga, Nabugabo, Albert, Edward
and George; the Semliki river; the Victoria Nile; Murchison and Albert Niles;
Elsewhere: Nile; Nigeria; Gold Coast; Lake Tanganyika.
Daget (1954) recognizes a distinct sub-species, Bagrus d. niger for fishes from
the Upper Niger.

Bagrus bayad (Forsk.) 1775. (Fig. 40b)
Native names: Lanya (Lunyoro and Alur); Oreko (Jonam).
Description: Similar to Bagrus docmac, but readily distinguished by its
narrower head, more slender body and by the position of the dorsal fin. Also,
whereas in B. docmac the dorsal surface of the head is smooth, in B. bayad it is
often ridged.
Coloration: Dark grey-black above, creamy-white below.
Size: Smaller than B. docmac; most individuals are about 40 cm. long.
Food: Predatory on small fishes and insect larvae (Worthington, 1929a).


FIG. 40b
Dorsal surface of the head of B. bayad. (Drawn by Barbara Williams.)

Habitat: Common in the deeper waters of Lake Albert (Worthington, op. cit.)
but also occurring inshore.
Breeding: Unknown.
Distribution: Uganda: Lake Albert. Elsewhere: Nile; Niger; Senegal; the
Chad basin.

Auchenoglanis Ginther 1865
Body moderately elongate, feebly compressed. Head massive, the heavily
sculptured bones clearly visible beneath the skin. Dorsal fin short, comprising
a spine and 7 or 8 rays. Adipose dorsal fin long. Only three pairs of unbranched
circum-oral barbels are present (the nasal barbels are absent).
One species occurs in Uganda.

Auchenoglanis occidentalis (Cuv. and Val.) 1840. (Fig. 41)
Native names: Bubu (Lunyoro, Lugungu, Alur); Orukwe (Jonam).
Description: Depth of body contained 3 to 5 times in the standard length,
length of head 2j to 31 times. Mouth small and terminal, the lips thickened and
papillose. Snout long and pointed, about half the length of the head. Maxillary
barbel reaching to the middle of the eye, or, in small fishes, as far as the posterior
border. Outer mandibular barbels as long as, or somewhat longer than, the
maxillary barbels, almost twice as long as the inner mandibular barbels. Dorsal
fin with I, 7 rays, the spine rough on its anterior face. Pectoral spine strong,
granulate (or even with a few low spines) on its outer face, strongly serrated on
the inner face.
Coloration: Olivaceous brown, darker above. Often with large, black, or
dark spots which extend onto the dorsal and caudal fins. Spotting is more intense
in small fishes, and is invariably present.
Size: Specimens one metre long have been recorded (see Daget, 1954) but in
Uganda the average length of large specimens is about 50 cm.


FIG. 41
Auchenoglanis occidentalis. (Drawn by Barbara Williams.)

Habitat: Inshore areas of Lake Albert, and in the Murchison Nile. Daget
(1954) records that in the Upper Niger, A. occidentalis is common in the
swamps of the flood plains.
Breeding: Virtually unknown in Uganda. Worthington (1929a) suggests that
breeding may take place in the Murchison Nile and around the shores of the
lake, but the evidence on which he based his opinion is rather slender. In the
Niger, Daget (op. cit.) says that breeding takes place during the flood seasons.
Food: Insect larvae, worms and small Crustacea (Worthington 1929a).
Economic importance: Slight; the species contributes to catches in seine nets
and 5-inch mesh-nets of fishermen on Lake Albert. Worthington (op. cit.)
records that most Africans refuse to eat 'Bubu' as they believe the flesh to be
poisonous. There is no evidence to support this notion.
Distribution: Uganda: Lake Albert, the Murchison and Albert Niles. Else-
where: Blue and White Nile; Lake Rudolf; Lake Chad; Senegal; Niger; Congo;
Lakes Mweru, Bangweulu and Tanganyika.

Family: Schilbeidae
Body scaleless, moderately elongate and strongly compressed. Dorsal fin
short, or, absent (present in the two genera occurring in Uganda), anal fin very
long. Generally four pairs of circum-oral barbels.
The Schilbeidae are found in the fresh waters of Africa and Asia. Two genera
occur in Uganda. Lake Edward is peculiar in that no members of this family
are present.

Adipose dorsal fin absent Schilbe
Adipose dorsal fin present Eutropius

Schilbe Cuvier 1817
Body strongly compressed. Dorsal fin very short, comprising a spine and 5 or
6 branched rays. No adipose dorsal fin. Anal fin long, extending from the vent
almost to the origin of the caudal fin. Four pairs of short, circum-oral barbels.
One species occurs in Uganda.

Schilbe mystus (Linn.) 1762. (Fig. 42)
Native names: Karakogere (Lunyoro); Nzere (Lake Victoria); English:
Description: The figure, together with the generic description given above,
will suffice to identify this species.

FIG. 42
Schilbe mystus. (Drawn by Barbara Williams.)

Coloration: Silver grey, somewhat darker above. The young are marked with
three longitudinal black bands, the upper and lower of which are continued onto
the caudal fin.
Size: A small species, rarely exceeding 30 cm. in length.
Food: Principally small fishes (Haplochromis spp.) although insect larvae are
also eaten (Graham, 1929; Worthington, 1929a; E.A.F.R.O.).
Habitat: Lake Victoria: Shallow inshore waters (E.A.F.R.O.; Graham, op.
cit.). Lake Kyoga: Open waters (Worthington, 1929a). Lake Albert: Probably
confined to inshore waters (E.A.F.R.O.).
Breeding: Unknown in Uganda. Records collected by E.A.F.R.O. suggest
that Schilbe mystus in Lake Victoria may breed during the rainy seasons. In the
Niger, the species spawns during the flood seasons. (Daget, 1954.)
Economic importance: Slight; seasonally high catches are made in Lake
Victoria. The flesh is very palatable.
Distribution: Uganda: Lakes Victoria, Kyoga, Nabugabo and Albert; the
Victoria Nile; Semliki river. Elsewhere: The Nile; rivers of west Africa; the
Chad basin; the Congo and Zambesi.

Eutropius Miiller and Troschel 1849
Closely resembles Schilbe, but differs in possessing a small adipose fin. One
species occurs in Uganda.

Eutropius niloticus (Riipp.) 1829
Native name: Taitai (Lunyoro, Lugungu, Alur and Jonam).
Description: Very similar to Schilbe mystus in all respects, including colora-
tion, but easily recognized by the presence of a small adipose dorsal fin situated
near the caudal fin.

Size: Somewhat larger than Schilbe mystus; large fishes average about 35 cm.
standard length.
Food: Small fishes and insect larvae (Worthington, 1929a).
Habitat: Bays and shallow inshore areas of Lake Albert; Worthington (op.
cit.) is of the opinion that in Lake Albert Eutropius occupies the ecological
niche held by Schilbe in the other large lakes. Little is known of its ecology.
Economic importance: Very slight. According to published accounts, the
flesh of Eutropius niloticus has a muddy flavour.
Distribution: Uganda: Lake Albert and the Albert Nile; Semliki river. Else-
where: Nile; the rivers of west Africa; the Chad basin.

Family: Clariidae
Body scaleless and elongate. Head bony, broad and flattened. Dorsal fin
long, without a spine. Anal fin long, extending to, or, almost to the caudal fin.
Adipose dorsal fin sometimes present. Four pairs of unbranched circum-oral
The Clariidae are widely distributed in Africa, Syria and south-east Asia.
Two genera occur in Uganda.

No adipose dorsal fin Clarias
Long adipose dorsal fin present Heterobranchus

Clarias Scopoli 1777
Head flat and broad. Body elongate. Eyes small. Dorsal fin long, extending
from slightly behind the head to the caudal fin, from which it is narrowly
separated. Anal fin long. No spine in the dorsal fin.
At least five species occur in Uganda. Like Barbus, the Clarias species of
Africa are in urgent need of taxonomic revision. Intra-specific variability is
high and many so-called species will undoubtedly prove to be merely geographi-
cal, or even individual, variants of a single, widespread species.
The hardiness of Clarias is almost legendary, and, although some of the
stories may be exaggerated, there is no doubt that Clarias can and do tolerate
conditions which would kill other fishes. That Clarias manage to survive in
foul, stagnant water is due mainly to their possessing an elaborate accessory
breathing organ which enables these fishes to utilize atmospheric oxygen. The
accessory organs are in the form of much branched bodies developed from two
of the gill arches on each side. These arborescent bodies are contained within
an expansion of the gill-chamber, which is lined with a highly vascular skin and
protected by the bones of the skull. The entrance to the chamber is guarded by
four fan-like processes from the gill-arches. Both the tissue covering the
arborescent organ and the skin lining the suprabranchial chamber have the same
microscopical structure as the normal gill-filaments. It is across these surfaces
that oxygen is absorbed from the air. The gill-filaments are well-developed, but
it appears that Clarias is dependent on the additional supply of atmospheric
oxygen. Fishes prevented from gaining access to the surface are 'drowned'.

There are many records, from other parts of Africa, of Clarias moving from
one body of water to another, but, to the best of my knowledge, this has never
been reported from Uganda. When temporary streams and swamps dry out,
Clarias are reputed to burrow into the mud and remain there until the area is
again flooded. However, further study of this supposed aestivation is required.
Certainly in a damp environment Clarias can live out of water for over eighteen
hours (personal observations on C. mossambicus). It would be no exaggeration
to say that there is hardly any water in Uganda which is without a population
of Clarias. In many cases, new dams, far removed from a stream or pond, are
rapidly populated. The amphibious habits of Clarias may explain this mystery.

Pelvic fin bases about midway between the tip of the snout and the origin
of the caudal fin, or, slightly nearer the snout 1
Distance between the pelvic fin bases and the snout contained 11-2 times
in the distance between the bases of the fins and the origin of the caudal
fin 2
1 Vomerine tooth-band as broad as the band of premaxillary teeth
C. mossambicus2
Vomerine tooth-band from 1 to twice as broad as the band of pre-
maxillary teeth C. lazera3
2 Nasal barbels shorter than the head; maxillary barbels not reaching
the first dorsal fin-ray. Distance between the tip of the supraoccipital
process and the first dorsal ray contained 1+-2) times in the head
length C. carsoni4
Nasal barbels as long as, or, slightly longer than the head; maxillary
barbels reaching the first dorsal fin ray or somewhat beyond. Distance
between the tip of the supraoccipital process and the first dorsal ray
contained 23-4 times in the head length C. alluaudi group

Clarias mossambicus Peters 1852. (Fig. 43)
Native names: Mali (Lunyoro, Luruli, Lunyara, Lukenyi); Nyaki (Ludope);
Twang (Lango); Mali or Kasonzi (Lake Victoria). English: Cat-fish, Barbel,
Description: Depth of body contained 6-8 times in the standard length,
length of head 3-3+ times. Head width contained 11 to 1i times in the head
length; upper surface coarsely granulate in adult fishes, smooth in the young.
Eyes small. Teeth on the premaxillae and lower jaw small, fine and arranged in
several rows. Vomerine tooth-band composed of several rows of granular teeth;
width of this band equal to that of the premaxillary tooth band. Nasal barbels
from 1-1 as long as the head in fishes longer than 12 cm., and from I to A of the
head length in smaller individuals; maxillary barbels rarely shorter than the
head, usually somewhat longer and reaching to a point midway between the
2 See text for note on C. anguillaris.
3 See text for note on C. moor.
4 See text for note on the closely related C. phillipsi.

origin of the dorsal fin and the insertion of the pelvic fins (as is usual in Clarias,
the barbels are proportionately longer in small fishes). Outer mandibular bar-
bels longer than the inner pair. Gill rakers long, fine and close set, their number
increasing with the size of the fish; from 25 (in small fishes) to 100 on the first
arch. Dorsal fin with 62-80 rays, anal fin with 50-65; both fins distinct from the
caudal fin.

FIG. 43
Clarias mossambicus. (Drawn by Barbara Williams.)

Coloration: Generally dark greyish-black above, creamy-white below; a
fairly distinct black longitudinal band on each side of the ventral surface of the
head. Young fishes less than 9 cm. long have a similar coloration except that the
black bands on the head are absent; larger young (10-30 cm.) are mottled, grey-
khaki above. This marbled coloration may persist in larger individuals. It is
likely that coloration varies with habitat; fishes kept in an aquarium with a
sand bottom are always very lightly coloured.
Size: In Lake Victoria, C. mossambicus reaches a large size. Most individuals
caught by African fishermen are between 80 and 90 cm. long, although fishes
120 cm. long are occasionally caught. In Lake Kyoga, on the other hand, the
species is smaller. Whereas in Lake Victoria sexually mature fishes, of either
sex, less than 50 cm. long are rare, in Lake Kyoga sexually mature females of
25-40 cm. are common. Fishes weighing 75 lb. have been recorded (Uganda
Government, 1948).
Food: Omnivorous, but predominantly a predator on small fishes (particu-
larly Haplochromis). Other food includes insect larvae, molluscs and plants.
The very fine and numerous gill-rakers enable the fish to filter the microscopic
organisms of the phyto- and zooplankton from the water; however, only a few
fishes have been recorded as feeding exclusively on plankton. Clarias living near
cormorant breeding colonies feed almost exclusively on regurgitated fish dropped
by the birds. Fledglings and unhatched eggs are also eaten. (E.A.F.R.O.)
Small Clarias (less than 5 cm. long) feed mainly on insect larvae, but the fish-
eating habit is developed by the time the young fish are 6 cm. long. (E.A.F.R.O.)
Habitat: Lake Victoria: Widely distributed, but more common in shallow,
inshore waters; also found in the marginal water-lily and papyrus swamps..
During cormorant breeding seasons Clarias are particularly abundant around
the islands on which the birds breed.
Lake Kyoga: Open water and in the water-lily swamps.
Lake Nabugabo: Ubiquitous.
Clarias mossambicus is also common in the major affluent rivers of Lake
Victoria, and in the Victoria Nile.

Breeding: In Lake Victoria, C. mossambicus is known to breed in small,
temporary streams which flow into the lake during the rainy seasons (Green-
wood, 1955). Although these streams may be connected with the lake for some
months, it seems that breeding fish run up on one or two nights only, in April
and again in December.
Clarias eggs are small (about 2 mm. in diameter) and are attached by an
adhesive disc to plants and debris in the bottom of the stream. Early develop-
ment is rapid, and the young fishes hatch within thirty-six hours after fertiliza-
tion. Despite their small size, larval Clarias are remarkably hardy and can
withstand considerable and rapid change in temperature; also, they are able to
utilize atmospheric oxygen at an early age, even though the supra-branchial
organ is not fully developed. The young remain in the stream for about six
weeks and then apparently swim down to the lake. During their first year, or
even two, the young reappear in the streams whenever these are connected with
the lake.
Pitman (in litt.) has described an interesting invasion by large Clarias (proba-
bly C. mossambicus) of temporarily inundated swampy ground south of the
western end of Lake Nakavali. The invasion took place in October, and,
although there is no positive evidence, it is probable that these fishes were
breeding in the temporary swamp.
Economic importance: Slight in Lakes Victoria and Kyoga. During spawning
runs large numbers of Clarias are speared by Africans, but this is only a
temporary affair. The flesh of Clarias makes good eating and is readily accepted
by most tribes.
Distribution: Uganda: Lakes Victoria, Kyoga, Nabugabo, Kachira, Nakavali
and Kijanebalola; the major affluent rivers of Lake Victoria and the Victoria
Nile. Elsewhere: Rivers of East Africa; Lake Tsana, at the headwaters of the
Blue Nile; the Zambesi and Lake Tanganyika.
Clarias anguillaris (Linn.), a Nilotic species, is thought to occur in Lake
Victoria. Externally, this species resembles C. mossambicus. It differs, however,
in having fewer gill-rakers in the adult (20-30 on the first gill-arch) and having
the vomerine teeth pointed and not granular. No specimens of C. anguillaris
have been caught by E.A.F.R.O., nor were any taken by Michael Graham's
survey (Graham, 1929). It is likely, therefore, that the record of this species from
Lake Victoria was based on the misidentification of a specimen of C. mossam-

Clarias lazera Cuv. and Val. 1840
Native names: Mali (Lunyoro and in general use); Nyai (Jonam).
Description: Similar to Clarias mossambicus; the principal difference lies in
the broader band of vomerine teeth (1 to twice as broad as the band of pre-
maxillary teeth).
Size: Specimens 140 cm. long have been recorded.
Food: Omnivorous, but mainly a predator on small fishes (Haplochromis and
young Tilapia) (Worthington, 1932b; Poll, 1939). Poll (op. cit.) records a

Tilapia 35 cm. long recovered from the stomach of a large C. lazera. Insect
larvae, molluscs, planktonic organisms and water-weed are also eaten by this
species. (Worthington, op. cit.)
Habitat: Lake Albert: Shallow water (Worthington, 1929a). Lake Edward:
Widely distributed, but especially common at the mouths of rivers and near
papyrus swamps (Worthington, 1929a; Poll and Damas, 1939). Lake George:
Ubiquitous (Worthington, 1932b).
Breeding: Unknown in Uganda.
Economic importance: Of little importance in the fisheries of Lakes Albert
and Edward, but ranks fourth amongst the commercially important fishes of
Lake George (Uganda Government).
Distribution: Uganda: Lakes Edward, George and Albert; the Semliki river;
Kirima river. Elsewhere: The Nile; rivers of west Africa, including the Congo;
the Chad basin; Syria and Palestine.
Clarias moori Boulenger 1901, known only from a single specimen from
Lake Edward, is almost certainly an aberrant C. lazera. According to the
description the principal difference between the two species is that the vomerine
tooth-band of C. moori is narrower than the premaxillary band, and the vomer-
ine teeth are conical and not granular.

Clarias carsoni Boulenger 1903
Native names: Usually included in 'Mali' the general term for cat-fishes; also
known as Nsonzi (Lunyoro, Lukenyi).
Description: Depth of body contained 6-9 times in the standard length, length
of head 4-5 times. Head smooth above, its width contained 11-1 times in the
head length. Eye contained 31-5 times in the head length. Nasal barbels shorter
than the head; maxillary barbels shorter than the head, usually reaching the
gill-opening. Outer mandibulars 11-1- times longer than the inner pair. Gill-
rakers coarse, 10-12 on the first gill-arch. Dorsal fin with 62-80 rays; the distance
between the first ray and the supraoccipital process is contained 11-2 times in
the head length. Anal fin with 55-65 rays. Pelvic fins situated markedly nearer
the snout than the origin of the caudal fin.
Coloration: Grey-black, lighter below.
Size: Rarely exceeding 35 cm.; most individuals are about 20 cm. long.
Food: Principally insectivorous (Worthington, 1932b; E.A.F.R.O.), although
larger individuals are piscivorous.
Breeding: Unknown.
Habitat: Lake Victoria: Marginal water-lily swamps, and in papyrus swamps.
Lake Edward: Papyrus fringe of Lake George and the Kasinga Channel. Lake
Bunyoni: Marginal weed-beds.
Economic importance: Only in Lake Bunyoni (where the species was intro-
duced) does C. carsoni contribute to a commercial fishery. It is frequently
caught in dams.
Distribution: Widespread in the streams (both temporary and permanent) of
Uganda, and in Lakes Victoria, Kyoga and Edward. Elsewhere: Lake Mokoto
(near Lake Kivu).

Worthington (1929a) records Clarias phillipsi Norman 1925 from Lake
Kyoga. This species closely resembles C. carsoni but is differentiated by its
slightly longer maxillary barbels (from as long as, to 1 times longer than, the
head); it was previously known from the Kigezi district of south-west Uganda
and Lake Kivu. The specific status of C. phillipsi is doubtful; it may represent
a variant of C. carsoni.

Clarias alluaudi group. (Fig. 44)
Native names: Nsonzi (Lunyoro, Luruli, Lukenyi); Mali (Lake Victoria).
Description: Clarias alluaudi Boulenger 1906, C. werneri Boulenger 1906,
and C. eupogon Norman 1928, are included in this description. The three
species are distinguished by the length of their nasal and maxillary barbels.
However, these characters show considerable individual variability and speci-
mens intermediate between C. werneri and C. alluaudi are common. C. eupogon
has the longest barbels of the trio (nasal barbels twice, and maxillary barbels
three times, longer than the head). It is known only from one specimen caught in
Lake Victoria. Thus, more specimens are required before it will be possible to
determine whether C. eupogon is a distinct species or not. There can be little
doubt that C. werneri and C. alluaudi are, in fact, one species (see also Poll,
Depth of body contained 6-8 times in the standard length, length of head
4-5 times. Head smooth above, its width contained 11-1 times in its length.
Eye small. Nasal barbels longer than the head, variable in length, reaching at
least to a point mid-way between the origin of the dorsal fin and the supraocci-
pital process, but often extending to well beyond the first dorsal fin rays. Maxil-
lary barbels usually longer than the head and extending to the first dorsal rays
or somewhat beyond. Outer mandibular barbels longer (1l-1j) than the inner
pair. 12-20 gill-rakers on the first gill-arch. Dorsal fin with 65-90 rays, its
distance from the supraoccipital process contained 2J-4 times in the head length.
Anal fin with 55-72 rays. Both the dorsal and anal fins are separated from the
caudal by a very narrow space.

FIG. 44
Clarias alluaudi. (Drawn by Barbara Williams.)

Coloration: Dark khaki, occasionally greyish-black; lighter below.
Size: Small fishes, rarely exceeding 15 cm. in length; most adult individuals
are about 11 cm. long.
Food: Mainly insectivorous (E.A.F.R.O.).

Habitat: Lake Victoria: marginal grass, water-lily and papyrus swamps.
Habitats in Lakes Kyoga and Edward are probably similar. According to
Worthington (1929a), C. alluaudi and C. werneri reach a larger size (23-25 cm.)
in Lake Kyoga than in Lake Victoria.
Breeding: Unknown.
Economic importance: Nil.
Distribution: Uganda: Lakes Victoria, Kyoga, Nabugabo, Edward, George,
Kachira, Nakavali, Kijanebalola; the Victoria Nile; rivers flowing into Lake
Victoria; streams and rivers of Uganda. Not recorded from any other area.

Heterobranchus Geoffroy Saint-Hilaire 1809
Differs from Clarias mainly in possessing an adipose dorsal fin. One species
occurs in Uganda.

Heterobranchus longifilis Cuv. and Val. 1840. (Fig. 45)
Native name: Unknown.
Description: The generic description, together with the figure, will serve to
identify this species.

FIG. 45
Helerobranchus longifilis. (Drawn by Barbara Williams after Boulenger.)

Coloration: Dark greyish-black above, lighter below.
Ecology: No data are available for this species in Uganda, where it is rarely
encountered. Worthington (1929a) records H. longifilis only from the turbulent
waters below the Murchison Falls.
Until very recently (Hulot, 1956), H. longifilis had not been recorded from
Lake Edward, although there were accounts of a Heterobranchus-like fish
having been seen by fishermen. However, there can no longer be any doubt that
this fish does occur in the lake (Hulot, 1956).
Distribution: Uganda: Murchison Nile; Lake Edward. Elsewhere: Nile;
Congo; Niger; Gambia; Zambesi; Lakes Tanganyika and Chad.

Family: Mochocidae
Body scaleless. Head heavily 'armoured'. Dorsal fin short, with a well-
developed spine. Adipose dorsal fin long. Anal fin short. Mandibular barbels
One genus occurs in Uganda. The Mochocidae are absent from Lake Edward.

Synodontis Cuvier 1817
Dorsal fin short, comprising a strong spine and 6 or 7 branched rays. Adipose
dorsal fin large. Pectoral fin with a strong spine. Three pairs of barbels (the
nasal barbels are absent), of which the two mandibular pairs are branched.
The pectoral spine is articulated by means of a complicated joint, so arranged
that the spine can be firmly locked into its erected position. The serrated spines
can inflict a painful wound, and these fishes should be handled with great care,
even when dead. Although the spines seem to provide a powerful defensive
mechanism, young Synodontis are eaten by predatory fishes and birds, whilst
large adults have been found in crocodile stomachs (E.A.F.R.O., 1952).
Synodontis are sometimes referred to as 'talking fish' because of the grunts
and squeaks produced when the fish is taken from the water. These sounds are
produced by movements of the pectoral fin spine.

Outer mandibular barbels with slender branches 1
Outer mandibular barbels with short, tuberculate branches S. nigrita
1 16-21 mandibular teeth; humeral process 1-11 times longer than deep
S. victoria
23-36 mandibular teeth; humeral process twice as long as deep S. schall
30-48 mandibular teeth; humeral process 1-l1 times longer than
deep S. frontosus
32-54 mandibular teeth; humeral process 11 times, to twice, as long as
deep S. afro-fischeri

Synodontis nigrita Cuv. and Val. 1840
Native name: Unknown.
Description: Depth of body contained 3-31 times in the standard length.
30-35 mandibular teeth. Maxillary barbels with broad membranes proximally;
extending to about the middle of the humeral process. The internal and external
mandibular barbels with short, tuberculate branches. Humeral process 1 times
longer than deep. Dorsal fin with a strong spine, serrated on its posterior face,
and 7 or 8 branched rays; distance between the last dorsal ray and the origin of
the adipose fin contained 1-11 times in the length of the adipose fin. Pectoral
spines weakly serrated on the outer face, strongly serrated on the inner.
Coloration: Dark brown to blackish, often with irregular black spots. Caudal
fin spotted. The young are somewhat lighter in colour and covered with black
spots and blotches, which extend onto all the fins; the caudal fin has from 4-6
sinuous, black bars.

Food: Unknown for Uganda fishes. Daget (1954) records plant debris and
insects for specimens from the Niger.
Habitat: Unknown in Lake Albert. In the Upper Niger, the species is found,
in swamps, but not in the river itself (Daget, op. cit.). S. nigrita is apparently
uncommon in Lake Albert. A number of young were collected by E.A.F.R.O.
in the Semliki delta area.
Distribution: Uganda: Lake Albert. Elsewhere: Nile; Niger; Senegal, Gambiat

Synodontis victoria Boulenger 1906
Native names: Nkolongo (Lake Victoria); Kigogole (Lunyoro and Luruli);
Takutaku (Lunyara); Kaitampio (Lukenyi); Okoke (Ludope); Gologolo (Lango).
Description: Depth of body contained 33-41 times in the standard length.
16-21 mandibular teeth. Maxillary barbels reaching the posterior tip of the
humeral process, or occasionally, somewhat beyond. Mandibular barbels with
slender branches. Depth of humeral process contained once to 1 times in its
length. Dorsal fin consisting of a spine, weakly serrated on its posterior face,
and 6 or 7 branched rays. Distance between the last dorsal ray and the origin
of the adipose fin contained 11-2 times in the length of the adipose fin. Pectoral
spine moderately serrated on its outer face in small fishes and only feebly so in
larger individuals; strongly serrated on the inner face.
Coloration: Greyish-silver, with dark spots of variable size. Caudal fin
occasionally spotted. Coloration of juveniles unknown.
Size: Rarely exceeding 35 cm. standard length. Sexually mature females
11 cm. long are known from Lake Victoria. Worthington (1929a) is of the
opinion that in Lake Kyoga the species is smaller than in Victoria.
Food: Mollusca (particularly bivalves) and insect larvae (Graham 1929;
Worthington 1929a and E.A.F.R.O.). It seems likely that S. victoria is a facul-
tative feeder, and that the diet varies both with the habitat of the fish and the
abundance of the food organisms.
Habitat: Lake Victoria: Little precise information is available. S. victoria
is apparently most common in water more than 40 feet deep, although the
species does occur in shallow water (E.A.F.R.O.; Graham, 1929). Lake Kyoga:
Almost completely restricted to the open waters (Worthington, 1929a).
Breeding: Unknown.
Economic importance: Slight.
Distribution: Known only from Lake Victoria and its affluent rivers, Lake
Kyoga and the Victoria Nile.

Synodontis schall (Bloch-Schneider) 1801. (Fig. 46)
Native names: Wahrindi (Lunyoro, Lugungu, and in general use); Kamuduli
(Alur); Lungu (Jonam).
Description: Depth of body contained 3-4 times in the standard length.
23-35 mandibular teeth. Maxillary barbels reaching to a point mid-way along
the humeral process, or, to the posterior tip of this process. Outer mandibular
barbels with slender branches, the inner pair with somewhat stouter branches.

Depth of the humeral process contained 11-2 times in its length. Dorsal fin con-
sisting of a spine (granulated on its anterior face, weakly serrated behind) and
6 or 7 branched rays. Distance between the adipose dorsal fin and the last dorsal
ray contained 21-4 times in the length of the adipose fin. Pectoral spine granu-
late on its anterior face, strongly serrated on the posterior face.

6 cm.
FIG. 46
Synodontis schall. (Drawn by Barbara Williams.)

Coloration: Olivaceous above, lighter (somewhat creamy) below; irregular
spots sometimes occur on the body. Young fishes lighter, marbled by yellowish,
irregular and sinuous bands.
Size: Attains a length of over 40 cm. according to Sandon (1950) and Worth-
ington (1929a).
Food: Predominantly molluscs, but also insect larvae and even small fishes
(Worthington, 1929a).
Habitat: Lake Albert: Deep, open waters and also in shallower water nearer
the shore (Worthington, op. cit.; Uganda Government; E.A.F.R.O.).
Breeding: Unknown.
Distribution: Uganda: Lake Albert, the Albert and Murchison Niles; Semliki
river. Elsewhere: Nile; Lake Rudolf; Lake Chad; Niger; Senegal.

Synodontis frontosus Vaill. 1895. (Fig. 47)
Native name: Kwoke (Lunyoro, Alur, Jonam).
Description: Depth of body contained 3-4 times in the standard length. 30-48
mandibular teeth. Maxillary barbels usually reaching the posterior tip of the
humeral process, but occasionally only as far as its origin. Outer mandibular
barbels with slender branches, inner pair with shorter, tuberculate branches.
Depth of humeral process contained 1}-1 times in its length. Dorsal fin con-
sisting of a spine, smooth on its anterior face, strongly serrated on its posterior
face, and 7 branched rays. Distance between the last dorsal ray and the origin
of the adipose fin contained 2-4 times in the length of the adipose fin. Pectoral
spine with a strongly serrated inner face; smooth anteriorly in fishes more than
20 cm. long, but serrated in smaller individuals.


FIo. 47
Synodontis frontosus. (Drawn by Barbara Williams.)

Coloration: Dark olivaceous or brown, peppered with small black spots.
Size: Attains a length of 35 cm. in Lake Albert.
Food: Little information available; from Worthington's data (1929a) S. schall
is a bottom feeder.
Habitat: Deep (120 feet) open waters and shallow inshore areas of Lake
Albert (Worthington, 1929a; E.A.F.R.O.).
Breeding: Unknown.
Economic importance: Negligible.
Distribution: Uganda: Lake Albert; the Albert and Murchison Niles. Else-
where: The White Nile; Omo river; Lake Rudolf.

Synodontis afro-fischeri Hilgend. 1888. (Fig. 48)
Native name: Unknown.
Description: Depth of body contained 31-4 times in the standard length.
32-54 mandibular teeth. Maxillary barbels reaching almost to the origin of the
pelvic fins in some fishes, and to the posterior tip of the humeral process in
others. Outer mandibular barbels with slender branches, inner pair with shorter,
but slender, branches. Depth of humeral process contained 1-2 times in its
length. Dorsal fin consisting of a spine, smooth anteriorly and serrated on its
posterior face, and 7 branched rays. Distance between the last dorsal ray and
the origin of the adipose fin contained once to li times in the length of the
adipose fin. Pectoral spine strongly serrated on its anterior face, very strongly
serrated on the posterior face.
Coloration: Marbled yellowish-brown. The extent of the marbling is
extremely variable, and uniformly brown fishes have been seen. In many fishes,
scattered dark spots occur on the body.
Size: A small species rarely exceeding 15 cm. standard length.
Food: Molluscs and insect larvae (E.A.F.R.O.).


2 cm.
FIG. 48
Synodontis afro-fischeri. (Drawn by Barbara Williams.)
Habitat: There are few data available from either Lake Victoria or Lake
Kyoga. It seems likely that in Lake Victoria, Synodontis afro-fischeri replaces
S. victoria in the shallow waters. The species has not been caught in water over
60 feet deep. It is common in the marginal vegetation of the Kagera river.
Breeding: Unknown.
Economic importance: Nil.
Distribution: Known only from Lakes Victoria, Nabugabo and Kyoga; the
affluent rivers of Lake Victoria, and the Victoria Nile.

Family: Amphillidae
Body scaleless. Head and anterior part of the body somewhat flattened.
Dorsal and anal fins short, the former without a spine. Adipose dorsal fin short.
Outer ray of pelvic fin thickened but not bony. Three pairs of unbranched
circum-oral barbels (the nasal barbels absent).
This family of small cat-fishes is confined to Africa. A single genus occurs in
Amphilius Giinther 1864
Diagnostic characters as for the family. Only one species is recorded from

Amphilius jacksoni Boulenger 1912. (Fig. 49)
Native name: Unknown.
Description: Depth of body contained 61-7 times in the standard length,
length of head 42-5 times. Head flattened, slightly longer than broad. Snout
rounded. Eye small, its diameter contained 5-6 times in the length of the head.
Posterior nostril nearer the eye than the end of the snout. Maxillary barbels
shorter than the head, reaching to slightly beyond the first, thickened, pectoral
ray. Dorsal fin with one unbranched and six branched rays. Pelvic fins situated
behind the last dorsal ray.


FIG. 49
Amphilius jacksoni. (Drawn by Barbara Williams.)

Coloration: Marbled and blotched yellowish-brown above, lighter and uni-
form below. Fins yellowish; caudal spotted; dorsal and anal fins with a dark
base and an interrupted black line near the tip.
Size: Small fishes, not known to exceed 10 cm. in length.
Food: Unknown.
Habitat: Fast flowing streams.
Distribution: Uganda: Hima river (flowing into Lake George); Aduka and
Agoye rivers (E.A.F.R.O., collected by P. Hainsworth). Elsewhere: Rutshuru
river (flowing into Lake Edward) Belgian Congo.

Family: Malapteruridae
Body scaleless, cylindrical and somewhat bloated in appearance. Head
covered with thick skin. Rayed dorsal fin absent; a moderately developed
adipose fin present and situated near the caudal fin. Anal fin short, situated
below the adipose dorsal fin. Three pairs of unbranched circum-oral barbels
(nasal barbels absent). A well-developed electric organ encases most of the
This family is only found in Africa, where it is represented by a single genus
and species.

Malapterurus electricus (Gmelin) 1789. (Fig. 50)
Native names: Ntera (Lunyoro); Adinga (Alur); English: Electric cat-fish.
The Arabic name, 'Raad', meaning thunder, is singularly appropriate.
Description: The diagnostic characters given above, together with the figure,
will serve to identify this altogether repulsive fish. The electric organ is capable
of giving a powerful shock and the fish should be handled with great care,
preferably through some non-conducting medium. Previously, the electric organ
of Malapterurus was thought to be developed from certain elements in the skin.
Recent research (Johnels, 1956) strongly suggests, however, that the organ, like
that of other electric fishes (see Chapter II), is derived from modified muscle-
fibres. The function of the electric organ is unknown. It undoubtedly serves as
a powerful means of defence and may also stun and immobilize the fish's prey.
Despite the unpleasant appearance of M. electricus, its flesh is reputed to make
good eating.


3 cm.

FIo. 50
Malapterurus electricus. (Drawn by Barbara Williams.)

Coloration: Greyish-silver above, dead-white below, the body and fins covered
with irregular and often large black blotches. The caudal and anal fins are
outlined with a pinkish band.
Size: Specimens 120 cm. long are known from the Niger (Daget, 1954) but no
data are available from Uganda.
Ecology: No information has been collected for this species in Uganda. In
the Malagarasi river (Tanganyika Territory) M. electricus was caught in the
reed-beds flanking the fast-flowing sections of the river. These fishes had been
feeding on other fish.
Distribution: Uganda: Lake Albert. Elsewhere: The whole Nile system; Chad
basin; Lake Rudolf; rivers of West Africa; the Zambesi; Lake Tanganyika and
the Malagarasi river.

Distribution of Barbus species in Uganda

LAKES Victoria Kyoga and Albert Elsewhere
Barbus altianalis P P P Victoria Nile
Barbus bynni P Murchison and Albert Nile
Barbus amphigramma P Streams and rivers flowing
into Victoria and Kyoga
Barbus portali P Mpanga, Waisoke, Kirima
Hima and Mongiro rivers
Barbus minchini .. P P
Barbus kersteni P Lakes Bunyoni and Mutanda
Barbus nummifer .. P Affluent streams of Victoria
Barbus somereni Ruwenzori streams below
5,500 ft.
Barbus prince P P ? Semliki river
Barbus neglectus .. Albert Nile
Barbus doggetti P
Barbus magdalenae P Lake Nabugabo; streams
flowing into Victoria
Barbus apleurogramma P Streams flowing into Lake


Distribution of Siluroid fishes in Uganda

LAKES Victoria Kyoga and Albert Elsewhere
Bagrus docmac P P P P Lake Nabugabo, Semliki
river, Albert, Victoria and
Murchison Niles
Bagrus bayad .. P
Auchenoglanis occidentalis P Murchison and Albert Niles
Schilbe mystus P P P Lake Nabugabo, Victoria
Nile, Semliki river
Eutropius niloticus P Albert Nile; Semliki river
Clarias mossambicus .. P P Lakes Nabugabo; the Koki
lakes, Rivers flowing into
Clarias lazera P P Semliki and Kirima rivers
Clarias carsoni .. P P P Widespread in streams
Clarias alluaudi group P P P Lake Nabugabo; the Koki
lakes; widespread in streams
Synodontis nigrita P Semliki river
Synodontis victoria P P Victoria Nile
Synodontis schall .. P Albert and Murchison Niles
Synodontisfrontosus .. P Albert and Murchison Niles
Synodontis afro-fischeri P P Lake Nabugabo, Victoria
Amphiliusjacksoni Hima, Aduka and Agoye
Malapterurus electricus P ? Albert Nile
Heterobranchus longifilis .. P ? Murchison Nile


Gilchrist, J. D. F., and Thompson, W. W. (1913). The fresh water fishes of South
Africa. Ann. S.A. Mus., 2, 321-463.
Greenwood, P. H. (1955). Reproduction in the cat-fish Clarias mossambicus Peters.
Nature, 176, 516.
Hulot, A. (1956). Apercu sur la question de la pIche industrielle aux lacs Kivu,
Edouard et Albert. Bull. Agric. Congo Belge, 47.
Johnels, A. G. (1956). On the origin of the electric organ in Malapterurus electricus.
Quart. J. micr. Sci., 97, 455-63.

(Presidential Address, 1956)'
Y OUR Excellency, Ladies and Gentlemen. I wish to speak tonight about
how local foods influence the work, wealth and happiness of the inhabitants
of Uganda and I will of necessity spend almost all my time talking about the
Africans, for Europeans and Asians eat their traditional foods and they are
almost always comfortably placed so that they suffer from no nutritional
deficiency. In fact they have usually a good height, due to excellent nutrition,
especially in protein, while they were growing in childhood, and some of them
have done so well in this calorie business, as I might call it, that they have
carried forward in their ledgers a balance in hand and these little accumulations
of capital have given them, what might be called euphemistically, a comfortable
My problem with regard to the Africans is how to make what might appear to
be a dull subject an interesting theme and the first condition for success is that
it should never be too technical and certainly not too tedious. So I will meander
round my subject, never going into any matter fully, but trying always to put
the essentials of the matter in straightforward English, so that all may under-
stand. I will even take the liberty of telling a personal story. I hold that any
very important scientific discovery, as opposed to some minor technical
improvement, is always susceptible of a simple explanation and can be explained
if not to children, well certainly to schoolchildren.
I find I like a garden most enormously but on certain terms: at the end of a
long and tiring day I love to stroll round my garden, which has been created by
my wife's imagination and the energy of my various shamba boys. All this work
I admire, rather like a secondary creator in a garden of Eden saying "Let there
be flowers", but my only difficulty is, like that of Adam, when I am called upon
to do the naming. So many of the flowers have the most impossible names and
I find that I invariably ask the name, admire the flower and then forget the
name. But if I am bad at botanical names there is for me no difficulty over the
more exotic names given to nutritional diseases, yet these are as impossible as
any seedsman's catalogue. There is beri-beri, a term from the East Indies,
pellagra which like sherry is a product of Spain, scurvy which is old English
for a mean and paltry fellow; rickets has as doubtful an ancestry as some of
our Kampala dogs and then there is finally Kwashiorkor. This is the latest
addition to this family of nutritional diseases; it is the baby, if not the spoilt
baby, and lest some of you should think it unimportant I will give only two
quotations. Many of you have never heard of Kwashiorkor, it is the latest
addition to this group of diseases, it is in fact so new that it is almost not men-
tioned in any book of medicine, and in very few books even on nutrition. One
of my tasks tonight is to discuss how a group of workers in the tropical parts of
I Delivered 22 February 1956.

Africa, America and Asia have recognized and studied this disease in the last
20 years. Now for the two quotations. Professor Brock and Dr. Autret in their
published report to the World Health Organization in 1952 wrote: "Kwashiorkor
is the most widespread and severe nutritional disorder known to medical
science"; and the Tropical Diseases Bulletin in 1955 in reviewing our book2
published on this disease said "No other single disorder can be responsible for
so much ill-health and suffering in childhood-with the possible exception of
malaria. There is no doubt that Kwashiorkor is the most important children's
disease in the world today". If you asked me to explain why Kwashiorkor has
only recently been discovered I would say that children and adults in Europe
almost always get enough to eat, that is enough calories, and their foods are
rich in protein; they may however suffer from vitamin deficiencies causing
rickets, scurvy, and so on. Malnutrition in Europe means almost exclusively
vitamin deficiencies, there was almost no clear knowledge in those countries
until the end of the last war of the results of underfeeding of the body with
calories and protein. This knowledge emerged during the war and fused with
our nutritional problems in the tropics.
The common foods all contain a source of energy, called calories and nitrogen-
containing substances, called proteins, out of which all the body cells are built.
In discussing undernutrition we will describe what happens when the body is
short mainly of calories; we do not often call it calorie deficiency because
actually the diet is short of protein and all the vitamins, only calorie deficiency
is the dominant deficiency. Similarly Kwashiorkor is almost entirely protein
deficiency, and many of you will prefer to call it such. Have it as you like, it is
the difference between the gardener who says "Those lovely red flowers on
that leafless tree, I call it Kirikiti" and the botanist who says "I understand you
perfectly, but I prefer to call it Erythrina abyssinica".
It is difficult to say how many calories a person needs, although in a recent
scientific paper contributed to the East African Medical Journal, I discuss
this point as far as it affects Africans. A man needs calories in proportion to
his size, and his requirements can be divided into two portions which might be
compared to 'basic salary' and to an 'expatriation allowance'. His basic needs
are the calories needed to keep him alive and warm: for a 12 stone man these
may be 1700 calories a day. That is to say if this man lay on a bed of asphodel
and composed himself to quiet heavenly thoughts of peace he would need 1700
calories daily; but man is consumed with desires to do things, some pleasant,
like eating, talking and gently meandering around, others more painful and
deliberate like that of working-these may consume from 1000-1500 calories
and we might compare this to the expatriation allowance. For man has fallen
from this heavenly state of gracefully lying on a bed of asphodel and desires
to be full of activity on the earth. So he gets an expatriation allowance for
working. The total calories of basic plus those for work come to 2800-3500 for a
working man.
When, however, we come to ascertain how many calories Africans take we
find that very few of them consume the expected number. A few surveys in
Africa have demonstrated that some people get enough, but most seem to get too
2 Kwashiorkor, by H. C. Trowell, J. N. P. Davies and R. F. A. Dean. London. 1954.

few. A survey of Kampala labourers revealed in 1950 that they got apparently
only about 2000 yet they should have eaten at least 500 calories more. Is the
man hungry? Does he desire to save money? If so he will decrease his con-
sumption of food and then feeling tired he will decrease his activity to avoid
overspending his calories. Perhaps that is what we see when we gaze at a large
building being erected in Kampala and see that about half the porters are
sitting down. These wise business men are saving calories. Or should we put the
answer the other way: are these porters naturally so indolent, or is work so
unpleasant at tropical temperatures that they desire to work less and so they
need fewer calories and therefore they eat less? We cannot state which comes
first here in this circle of hunger and calories eaten; work, and calories con-
sumed. Break the circle anywhere and the wheel stops turning and the man
stops working and the mouth stops eating.
All over the tropical belt from India to West Africa and Salisbury to Cairo
we get records of low calorie intakes, not only in adults, but also in children;
thus a detailed survey by Dr. Dean and Miss Schwarz at one of our best local
boarding schools revealed low calorie intakes. The problem extends even into
childhood and Professor Holmes did a quantitative survey of Baganda toddlers
aged about 2 years; those from peasant families took 750 calories, those from
upper class African homes took 1125 calories; British children of the same age
would take 1275 calories. The young peasant children from a very early age
take too little food and they are not often seen bouncing with activity as are
better fed African children, too often they are sitting or standing somewhat
passively, saving their calories. We should not and must not talk about these
children as starved, or neglected by their parents, or eating bad food, but quite
obviously on their low calorie intakes they cannot be as active as better fed
children, and when they grow up they may continue set in their ways of rather
low calorie intakes and rather low physical activity. If you grow maize on a
poor soil and with little moisture it will be short and spindly; if when it is fully
grown you saturate it with water and fertilizers it will not produce a good cob.
It is too late. Even if you sat some of our short stunted porters to dine with
kings they might not blossom into activity. That is our calorie problem, barely
comprehended, even in its scientific aspects, and often neglected by those who
think we have only a problem in protein. But protein is little use unless you
have enough calories.
As all this must seem rather depressing and is getting very technical I have
thought that before we debate the question of protein intakes I would discuss,
if you will forgive me, the question of how we ran across this disease of protein
malnutrition called Kwashiorkor. How does a medical scientist set about a
problem of this dimension? I cannot of course tell the full story, but I thought
that some aspects of it might be of interest. How was one of the major diseases
of mankind unravelled? Of course it has been the work of many persons, often
working in ignorance one of the other.
In the beginning as far as I am concerned there was no design about it. I
worked at the disease some four years before I ever heard the word Kwashiorkor.
No Council of Medical Research put it down on the agenda and directed an
enquiry into Kwashiorkor. These august bodies can direct activity into recog-

nized problems such as malaria, or sleeping sickness or even cancer. No, I was
pushed into Kwashiorkor by complete accident. At the beginning of my second
tour in Kenya in 1932 I was given the task of endeavouring to train African
men as nursing orderlies in Nairobi. I had just come back rather pleased with
myself as I had obtained a Membership, no one at that time thought about
medical specialists, but I did take a dim view of my departmental posting which
severed me from sick patients. Eventually after a protest I obtained some
clinical opportunities; no one wanted the children's ward in the old K.A.R.
hospital in Nairobi, so I was given it. In sullen mood I went into the corrugated
iron building; it was dark, it was full of bed lice, the children were all crying,
but many of them had brown hair although they were African babies and many
of them had swollen legs with oedema. On walking through the ward at night,
lit by one hurricane lantern, one would hear the low moan of the children dying;
we did not know why they were dying, but it was Kwashiorkor. So there, as far
as I am concerned, it all started. The 400-bedded hospital had one water tap,
during all my six years it never had a microscope, no African could talk English,
we could not even examine the urine, stool, or blood slide in the hospital, every-
thing went to a laboratory situated over a mile away. We only knew that when
the baby developed black spots on the skin it was going to die. Then one day I
surprised my wife by saying that I felt fairly certain I was dealing with a new
disease. Its outstanding sign was the brown hair, the swollen legs of oedema,
but this is where the trouble started. There were so many known causes for
oedema, all sorts of heart diseases, kidney diseases and blood diseases caused
oedema. Might not some recognized disease be there but missed? You could
miss a lot, believe me, in the old Ward 5 of the Nairobi K.A.R. Hospital. Might
it not be due, as the other doctors thought, to some of the worms and parasites
which were present so often in the body? So we dosed them of their worms and
they went on dying. At that time we were losing about three-quarters of our
cases of Kwashiorkor.
It seemed impossible to think that there was a new disease just in Nairobi.
and then I made my first serious mistake. So many children had got swollen
legs that I said I would limit my enquiry to those with black patches on the
skin, this I called 'X Disease'; those with only swollen legs I called 'Nutritional
Oedema'. So I had two diseases; I did not realize that children suffering from
protein malnutrition first of all stop growing, and most never get any worse, a
few develop the swollen legs and some of these get better but a very few get so
ill that they develop the black patches and then they always die. So I limited my
enquiry to those having black patches and I read all I could about any allied
At that time in Nairobi there was only a small medical library, so one could
not do much reading, but clearly this 'black patch disease' was not in the text-
books. The thing to do was to study it, try and understand what was occurring
in the body, and above all to try and find what organs were affected; this meant
getting post-mortems and examining as carefully as possible heart, kidney,
blood and so forth. Alas, there were no changes in the body at post-mortem,
the disease had nothing distinctive about it; the blood was a little thin, the liver
a bit yellow, the muscles a bit flabby. Now no disease can be recognized by the

medical profession as a whole until certain distinctive changes can be found at
death. After all you do die of it; but at post-mortem we would only find in my
cases a bit of pneumonia here, or a few round worms there, or some malaria,
and death was recorded as due to these diseases. I could not prove the presence
of my 'X disease'. I would go time and time again to the post-mortems, hoping
always to find the secret in the brain, or the lungs, or the endocrine glands, but
it was always disappointment. From 1932 until nearly 1944 we searched and
searched, but could find nothing. Then towards the end of the war it all came
like a flood.
From fairly early on I had the idea that this was some queer kind of malnutri-
tion, possibly linked to protein, as I said in my first meeting in 1934 and then
the first news we got that someone else was on the same track was a letter in a
medical journal saying that Dr. Cicely Williams had described this disease in
West Africa, "but she had got it all wrong", so the nutrition expert said in his
letter, for in his opinion the black patches were obviously those of pellagra.
She also had made the same mistake as myself, she limited her enquiry to those
with the patches on the skin and the expert said these were pellagrous. By that
time I had collected a lot of photographs of the disease and sections of the
tissues and I took them back to the expert in London in 1935. I was young and
foolish and I listened to the expert. I called the disease 'Infantile pellagra' in my
first big article, but he did not quite convince me, so I said it was a mixture of
pellagra and nutritional oedema. Thus I launched on the world a new disease
'Infantile pellagra'; it persists in swimming round, a whole book has been written
on it in South Africa. It is alas complete bogus, and this was shown almost at
once, for scientists in America during that fatal year of 1935 isolated a new
B group vitamin, nicotinic acid, which is curative in pellagra; they sent me some
of the first preparations and I remember the day when I unwrapped the big
brown bottle that was to cure my cases of 'Infantile pellagra'. I treated ten cases;
eight died. I could have wept.
By this time I had transferred from Kenya to Mulago Hospital, Kampala,
but I could not leave my problem behind. I detected many cases at Mulago
Hospital, but was distressed to find that some doctors in Uganda called the
disease congenital syphilis. By this time I felt that perhaps the disease was
allied to pellagra and that I should try some of the other lesser known vitamins,
mainly those of the B group. So I was always trying this vitamin and that
vitamin, or lots of liver extracts because these are rich in B vitamins. There are
now 12 separate B vitamins which are known to man; many of these I tried one
by one and sometimes the baby got better, and sometimes he died. I also tried
these vitamins on adults, many of whom also had oedema and anaemia; in fact
at that time I could not distinguish between the children and the adults. Both
had 'X disease'. I did not neglect protein. I gave cow's milk and whole liver and
all the known vitamins and still I would lose about 40 per cent of my children,
and still at post-mortem there was no cause of death. I must have been a great
trial to my colleagues, some of whom were still attracted by the ideas of con-
genital syphilis or tropical infections; I fear I was a nuisance to my wife and
family, but they bore with me. It is a strange experience to walk and work, to
eat and sleep, for 13 years with a great problem and on two occasions I have

shown unhealthy signs of mental strain, but my problem was this: What is
killing these children, why cannot we find anything wrong at death, and finally
why, oh why, if it is malnutrition, why can I cure only two-thirds of the cases if
I give them milk and all the known vitamins? Surely it cannot be a malnutri-
tional disease!
There is nothing more heartening to the lonely traveller than to find that others
are exploring his area of country, although it may be annoying to find they got
there first. Uganda at that time had no medical library in Kampala, and the
only news I had of any other serious explorers was of Cicely Williams. She
replied to my title of 'Infantile pellagra' by saying I was wrong, but she said
this in a nice long scientific article and she said she was calling X-Disease
"Kwashiorkor", because the Africans near Accra in the Gold Coast, where she
was working, called it thus. These Africans noticed that soon after a child was
weaned his hair became brown, sometimes his legs became swollen and he
developed black patches and died. This disease they called Kwashiorkor. So
Cicely and I engaged in a long series of scientific articles in which I said it must
be 'Infantile pellagra', even if the nicotonic acid would not cure the pellagra,
and she kept on calling it 'Kwashiorkor', and pointing out all the weaknesses in
my argument. This argument started in 1932 when I was in Kenya and she was
in the Gold Coast; it continued when I got transferred to Uganda in 1935, and
then she got transferred to Malaya in 1935, and then the Japanese very stupidly
interrupted this correspondence of ours by invading Malaya. Cicely Williams
was herself captured and underwent great deprivations. But silence and suffer-
ing brought us together and when I met her for the first time in 1945 in England
she was a walking skeleton from a prisoner of war camp and I had been medi-
cally boarded with insomnia. But this insomnia was all over the disease for I
had proved that Cicely Williams was right: it was a new disease; it was not
pellagra, it was due to protein deficiency; we had found the changes in the body
at post-mortem and we had found the cure. This was all due to continued
efforts on my part in collaboration with a growing circle of helpers and a happy
chapter of accidents.
In 1939 on my leave I took three months off to do a course of study. This was
somewhat phoney for almost every afternoon I was able to slip away into the
grand library of the Royal Society of Medicine. I was certain some others must
have my X-disease, but one could not consult any index on this disease and no
one knew anything about my 'Infantile pellagra' or Cicely Williams' 'Kwashior-
kor'. I rummaged round in obscure tropical journals and old books of paedia-
trics and then it slowly came out: Normet had described in 1926 a curious
disease in Indo-China which he called 'Annam swelling'. Castellanos had
described in 1935 'Pellagroid-beri-beri' in Cuba; then I got one or two references
in the old French literature. Alas, German and Spanish are closed books to me
but I could look at the illustrations in the journals of various tropical countries
and there I found in many places and under many names virtually the same
disease. Some of the earlier articles were still missing, but I took notes of them;
thus Dr. Dean was recently able to see the article written by Correa in 1906 in
Mexico. All this literature was sifted by me in one final blast directed at Malaya
and towards Cicely Williams and then she never replied and I knew she was

captured, for being a woman she would reply at once if it was at all possible.
So the war closed down on us and the children went on dying. One vitamin
after another of the B group and all forms of protein, especially liver, were
tried; still nearly 40 per cent died. Then in 1942 I started reading the work of a
young man, a professor of medicine called Himsworth in London. He was
doing experimental work on malnutrition in animals producing liver disease,
especially fatty livers. Now the liver was always fatty in Kwashiorkor, but this
we had completely disregarded as we knew that diarrhoea could make the liver
fatty and these children often had diarrhoea. So I started posting pieces of liver
to Himsworth, and then I got back letter after letter, for in the livers of my
cases Professor Himsworth saw clear evidence of malnutrition.
As far as I was concerned the disease now had got something distinctive in its
pathology and it could now be recognized as a disease on its own. Towards the
end of the war we were joined at Mulago by a young pathologist, Professor
Davies, and he argued, very correctly, that if the children were having diarrhoea
there must be something wrong in the intestines or adjacent organs. As we were
still losing over one third of our cases it was easy to get post-mortems and in the
first one he made a most exhaustive examination. All sorts of pieces were
examined and special stains employed, and very soon a severe degeneration was
seen by him in the pancreas. It was making no secretion, for its secretion is
itself made from protein. Now the pancreatic secretions are the most important
digestive juices and now at last it all began to make sense. It had taken 13 years,
but at last I thought that I knew why I was losing over one-third of my cases,
although I gave them milk and all the known vitamins; they were unable to
digest this food! This explained, I then thought, my outstanding problem which
was-How can you die so easily of a malnutritional disease although you are
given a perfect diet?
After this things started to happen and happen very quickly. Soon Sir Harold
Himsworth, as he is now called, was to become the virtual Director of Medical
Research in Great Britain and the Colonies by his post at the Medical Research
Council. All sorts of excellent colleagues such as Dr. Margaret Thompson,
Professor Eric Holmes, Dr. Margaret Stannier, Dr. Lehmann, Dr. Welbourn,
Dr. McFie, and finally Dr. Dean and Miss Schwarz have come; a special
M.R.C. unit has been set up at Mulago thanks to the generosity of the Uganda
Government and other bodies. An encouraging feature is the large amount of
research work done in other countries, especially Central and South America
and India. I will say little about these recent advances, they depend now on
selecting persons of the highest ability. My colleagues, Professor Davies and
Dr. Dean, and I, have finished a book on Kwashiorkor. But my child has grown
up and is leaving home, I no longer recognize him, nor he me. I cannot even
master all the literature pouring out on Kwashiorkor; in New York they are
analysing the amino-acids in the urine of Kwashiorkor, in Cape Town they are
trying pure amino-acids in treatment. I look at Kwashiorkor now and know I
can never understand it all.
There is general recognition that this disease is absent in most of the countries
of Europe and North America and Australia, that is to say in the advanced
countries in the temperate regions. Here the staple food is wheat, or some other

cereal, and these contain in proportion to their calories about twice as much
protein as the starchy staple crops of the tropics (plantains, potatoes, yams, with
cassava almost pure starch). In addition cattle are plentiful in temperate regions,
at least during the last two centuries, with much milk for children. This is far
more important than meat, fish or eggs. In the southern countries of Europe,
such as Italy, Spain and Greece, there is some Kwashiorkor in the poorer
children, this has been recognized for many years in Italy, but in Spain its
recognition is more recent. In Mexico the disease is very common, as it is in
most countries with a Mediterranean climate, especially among poorer social
groups, and thus it is very common in Egypt, parts of India, and almost cer-
tainly China and Japan, although reports from the last two countries are few
and few doctors there have yet heard about it. In India recognition is almost
limited to the last six years; it is very common in the poorer families of rice-
eating groups, for rice is poorer in protein than wheat. The disease is common
throughout tropical Africa and South Africa, it is rare among pastoral tribes or
those giving a good mixed diet to children, especially if this contains leguminous
vegetable seeds as these are all rich in protein. Thus beans, peas and various
lentils all protect against Kwashiorkor, so there is little of this disease in Kabale
and in the Northern Province of Uganda. These areas produce fine specimens
of humanity who tend to carry off too many prizes at the Protectorate sports
and have for many years been the backbone of the K.A.R. and Police. In Africa
the disease is less in drier areas showing much seasonal change of rainfall, for
there are cereals, leguminous crops and cattle in these areas; it is more common
in the moist tropics. There is a great deal of the disease in almost all countries
of Central America and much in South America.
It is generally agreed that African children while breast fed grow well; in
fact a good deal of evidence is accumulating at Mulago Hospital that these
babies grow more quickly than other races and develop mentally at a faster
rate than European babies. An African baby is often sitting up unsupported at
four months, and walking well long before he is a year old. This is a very
important point and it shows that given an equal chance Africans can do as well
if not better than other races. No reason is at present known for this superiority
of the African baby; is it due to heredity or excellent feeding or prolonged
contact physical and mental with the African mother? We do not know, but as
an amusing hypothesis we might postulate that it is the latter and that in twenty
years time you will see Bond Street in London full of babies strapped to their
mothers' back and breast fed, on demand, in special alcoves off the beauty
parlours. Somewhere in the second six months or in the second year of life in
the vast majority of African infants growth slackens and as this is about the
time of weaning, or at least when breast milk is no longer adequate, its cause
is, almost certainly, mostly dietetic. The child is almost certainly getting too little
protein and some of them would be getting too few calories. This precious
infant flower is nipped in the bud; the lively baby too often turns into the dull
toddler, his body weight increases slowly, he falls behind the normal standards
and his hair often becomes brown. Given more food, more calories and even
better still more protein, he grows well and becomes lively.
Let me conclude with one final hypothesis. It is a pure hypothesis but I think

a reasonable hypothesis and one day I must try and work it up for a scientific
journal. Why have the higher civilizations almost always emerged in temperate
regions or the sub-tropical regions, such as India, and Egypt, but almost never
in countries lying mainly within the true tropics? There must be some reason
for this, it cannot be just a coincidence, and yet of the countries lying entirely
within the tropical belt there were large civilizations only in Ceylon and certain
countries of Monsoon Asia, all of which had the cereal rice, and there was also
the Aztec civilization of America which had maize. With these exceptions, and
they were rather small exceptions, no great civilization arose within the tropical
regions of America, Africa or the East Indies. But civilizations have only
occurred where there have been dense populations, with the stimulus of environ-
ment and the differentiation of skills. This was impossible in pastoral nomadic
tribes, who might have an excellent diet and would produce outstanding indi-
viduals, as seen among the Arabs, ancient Jews, and possibly also the Somali,
Hima and Masai. Cities became possible where dried cereals and other seeds
such as beans were grown and could be stored. Only thus could the calorie
problem of the city be solved. In addition cereals are relatively rich in protein;
per 1000 calories wheat has 35 grams of protein, rice 20 grams, but matoke
12 grams and cassava only 5 grams. Now most tropical foods, at least those
grown in the moist tropics of Africa and America, are not cereals, they are
themselves moist foods which decompose in a few days and they are low in
protein. Asia had rice in the moist tropics and this was the basis of town life in
Asia. The drier parts of Africa had millet and sorghum, but these did not grow
well, neither could they be stored in the moist tropics. All these cereals when
fed to children, even in the absence of milk and meat, supplied larger amounts
of protein and allowed of more rapid growth. Whether they led to superiority
of mental ability we do not know, but some difference in totaf performance is
The tropics had also tropical diseases, but it should be remembered that these
seldom seem to affect adults adversely; malaria is largely a disease of children.
There was also much malaria and other tropical diseases in Egypt, Greece,
Rome, Assyria and India, all of which produced civilizations, but had good
cereals and also pulses, milk, meat and fish, at least for upper classes. Tropical
temperatures may have discouraged physical effort, which becomes unpleasant
in warm climates, but it was just as warm in many parts of Babylon and Cairo
as in East Africa.
Civilizations seldom arose in the tropics because they never solved the two
problems of how to store calories or how to get protein-rich food to the growing
children. This is now possible, not by labelling any tropical food bad, they are
all good and reflect the beneficence of creation, but by adding on other foods,
those which can be stored, easily cooked and which are rich in protein. So
tonight we have been talking about the life, health and happiness of all the
under-developed areas of the world, of all the places which seldom had an
equal chance. We are gambling for large stakes tonight; we are talking about
millions of people and yet for me it all started with babies crying in the night.
This has been a long search and I have been in good company, receiving the
knowledge of the chemist and bio-chemist, this communion of scholars within

which we live and move and have our being. And my final words are these: We
could never find it, it was always too big for us; it was always there like beauty
and love, but we could not see; we did not find it by any efforts of our own, but
in the end our eyes were lifted. You never discover anything in science; you
only are allowed to be among some of the first to gaze upon it and you are
silent, like a child, with wonder.

Brock, J. F., and Autret. (1952). Kwashiorkor in Africa. W.H.O. Monograph Series,
No. 8.
Davies, J. N. P. (1948). Lancet, 1, 317.
Gillman, J., and Gillman T. (1951). Perspectives in Human Nutrition. New York:
Grune and Stratton.
Trowell, H. C. (1937). Archives of Disease in Childhood, 2, 193.
- Davies, J. N. P., and Dean, R. F. A. (1954). Kwashiorkor. London: Edward
Williams, Cicely (1935). Lancet, 2, 1151.