THE GALL A IN NORTHERN SOMALILAND •>

I. M. LEWIS

Riassunto.- Da tempo£ statoprovatoche una volta i Galla occuparono

zonedel territorio ora somalo.II presente articolomostrache,primadei Somali,
essistanziavano in alcunipuntidi quellocheoggi£ il Protettorato britannicoe ne
ricostruiscel'allontanamentosottola spintadell'espansione dei gruppisomalinella
Somaliasettentrionale.
R£sum£. - Depuislongtemps il est prouv£ que les Galla occupaientdes par-
tiesdu payssomaliavantles Somali.Le present articlemontre qu'avantles Somali
ils 6taient£tablis dans certainsendroits de l'actuelProtectoratbritannique et suit
leurd£placement dansla Somalieseptentrionale sousla pressiondes groupesSomali
avacant.
Summary. - Therehas longbeenevidencethatthe Galla occupiedpartsof
Somaliland before theSomali. Thispapershowsthatbefore theSomalitheywere
stationed in partsof whatis todaythe BritishProtectorate and tracestheirwith-
drawalunderpressure fromtheexpanding Somalicommunities inNorthern Somaliland.

That the Galla occupied what is today centraland southernSomalia

prior to the Somali is securelyestablished. Dr. Cerulli has brilliantly
reconstructedthe sequence of Somali migrationswhich, fromperhaps
as early as the 12th century,began to bringa medleyof Somali groups
fromtheir earlier positionsin the north of Somaliland into the centre
and south at the expenseof the earlierGalla inhabitantswho weredriven
south-westwardsinto Ethiopia. In the 17th century the Galla were
forcedacross the Juba River, and by about 1910 the Dar5d Somali had
advanced as far south as the Tana River in Kenya where they were
halted by the establishmentof fixedgrazingareas a). That at an earlier
period the Galla also preceded the Somali as the inhabitantsof Nor-
thern Somaliland is not so well established, although it is suggested
by traditionswhich trace the originsof the Galla to NorthernSomali-

x) This paper is based on materialcollectedin Somalilandduringmy

tenureof a ColonialSocial Science ResearchCouncilFellowshipin Social
Anthropology.I am greatlyindebtedto the Councilfor theirgenerosity
and to the Government of the SomalilandProtectorate. For criticismof
an earlierdraftof this paper I am especiallygratefulto Mr. B. W. Andr-
zejewskiand to Dr. M. Pirone.
Le popolazionidella Somalia nella tradizione
2) See B. CERUiyiyi, storica
locale, "RendicontiR. Accademiadei Lincei(ScienzeMorali) 1926 pp. 150-
172; I. M. Lewis. Peoples of theHorn of Africa , London, 1955, pp. 45-8;
and CERUUJ,Somalia, ScrittiVari Editi ed Inediti,vol. I, Rome, 1958.

This content downloaded from 193.205.142.142 on Mon, 01 Feb 2016 09:56:47 UTC
All use subject to JSTOR Terms and Conditions
22 I. M. I/EWIS

land 1). Some evidence has also been furnishedthat the Galla formerly
occupied northernMijerteinia2) (in the northof what is today Somalia),
and also parts of the presentOgaden regionof Harar Provinceof Ethio-
pia 3). In this paper I adduce new evidence which to my mind makes
it almost certain that the Galla did in fact occupy much of northern
Somalilandpriorto|the Somali.

I.

Before presentingmy evidence, it is necessaryto referbrieflyto

a linguisticpoint which has been a source of much confusionin evalu-
ating local Somali traditionsconcerningthe Galla. In contemporary
NorthernSomali thereare two distinctlinguisticexpressionswhichhave
oftenbeen taken as the same. These are Galla, the name by whichthe
Galla people are known to the Amharas, Arabs, and others,and gal
(PI. galo, and with the article,galada) a pagan, infidel,or non-Muslim
in general4). In contemporarySomali these are separate and distinct
terms;and Somali are well aware that many of the Galla people, whom
they also referto collectivelyas Gallawi or if they cultivate, as Qoti),
are Muslims and thereforenot galo. In Ethiopia, the Amharic word
Galla has the sense of " subjects those under government5);and al-
though the Galla peoples are known collectivelyto outsiders by this
title, they do not apply it to themselves. Indeed, there is apparently
no collective name used by all the various Galla groups to designate
themselvesas a distinctculturalentity6). It has howeverbeen suggested
that this name, whichis firstrecordedin the 16th century7) and which

x)See G. W. B. HunTingford,TheGallaofEthiopia,London,1955,p. 19.

2) CERUI^I,Tradizionistoriche e monumenti , " Africa
della Migiurtinia
Italiana ", iv, 1931, pp. 153-169.
3) See M. Pirone, Leggendee tradizionistorichedei Somali Ogaden,
and Le popolazionidelVOgaden,in " Archivioper l'Antropologia e 1'Etnolo-
gia lxxxiv, 1954,pp. 119-143.
4) As is well-known, the wordgal is commonlyapplied by Somali to
Europeans, often withoutanydirectly derogatoryintention.In Northern So-
malilandgal in thiscontextis usedas widelyas the moreneutralwordfaranji.
5) B. W. Andrzejewski,personal communication .
8) The name Oromo which includesthe majority the Galla people
of
does not includethe Boran Galla of the NorthernProvinceof Kenya and
cannot thereforebe taken as an ethnicname for the whole Galla group.
B. W. Andrzejewski, personalcommunication.
7) It is apparentlyfirstrecordedin Bahrey'sHistoryoftheGalla (1593).
See C. F. Beckinghamand G. W. B. HunTingford,SomeRecordsof Ethio-
pia, 1^3-1646, London, 1954; The Galla of Ethiopia,cit., p. n.

This content downloaded from 193.205.142.142 on Mon, 01 Feb 2016 09:56:47 UTC
All use subject to JSTOR Terms and Conditions
 THE GAI&AIN NORTHERN
SOMAIJIyAND 23

Cerulli apparently erroneouslyrefersto as the " ethnic name of the

Galla has been adopted by Somali in the formgal, infidelor pagan,
referring to a time when the Somali had adopted Islam but while the
Galla were still non-Muslims1). But no secure evidence has yet been
adduced in proofof this theory,and the fact remainsthat in modern
NorthernSomali, at least, galo and Galla are quite distinct,although
this may of course have not been so in the past. .
It is necessaryto establish this linguisticpoint because the many
stone mounds,loosely referredto as " Galla graves with which Nor-
thern Somaliland is studded, have often been taken as evidence that
the Galla formerlyoccupied the regiona) when Somali usually attribute
them not specificallyto the Galla people but ratherto unspecifiednon-
Muslims or pagans (galo). Somali generallyregard them as the work
of the " people who were before" - an ambiguous descriptionwhich
may mean anything- and because they differin constructionfrom
contemporaryIslamic burials suggest that they are the work of non-
Muslims. However, a Radio Carbon analysis of wood fromone such
grave which I excavated indicates that some at least of these burials
are not morethan two hundredand fiftyyears old and probablycontain
Somali remains8). The traditionswhichI discussin this paper and exa-
mine in the lightof otherevidence referspecificallyto the Galla people
rather than to unidentifiednon-Muslims.

n.
RESIDUALGAIXAGROUPSAMONGST
THE NORTHERNSOMAU.

I begin my argumentby discussingthe many small pocketsof Galla

which still survive today in the British Somaliland Protectorateand
Ogaden region of Harar Province of Ethiopia.
In 1946, the total population of the " Reserved Area " of Ethiopia
(then under British Administration)was estimated at 300,000 souls,
of whom 5,000 were Amharas, Arabs, and Indians, mainly residentin
Jigjiga. Of the remainder,10,000 were Galla, mainly Akiso, and the

x) See Ceruuj, op. cit., 1931,p. 154.

a) See e. g. Huntingford,op. cit.,1955,p. 19; Ceruuj, op. cit.,1931,
p. 163; and N. Puccioni, Antropologia delleGentidella Somalia,
e Etnografia
vol. iii, Bologna, 1936, p. 121.
8) I hope to reporton the resultof theseexcavationsand the analysis
of humanremainsfromthe gravesin a laterarticleon the so-called" Galla
"
graves of NorthernSomaliland.

This content downloaded from 193.205.142.142 on Mon, 01 Feb 2016 09:56:47 UTC
All use subject to JSTOR Terms and Conditions
24 I. M. IvEWlS

rest Somali l). The Akiso are a segmentof the Barentu or Eastern Galla
who forma prominentelementin Harar Province of Ethiopia and who
extendto the east of the ancientcity of Harar in the directionofJigjiga
towards British Somaliland. Among the Somali Yabarre, Bartirre,
Geri, and Abaskul clans of the Darod clan-family2), and the Bursukof
the Dir clan-family, today residentin the Harar-Jigjiga regionand stret-
ching towards the east and south-eastto the Ogaden, live small groups
of Galla Waraitu, Waradayo, Asabo, and L,ehile. Furtherto the east
in the Ogaden amongstthe Somali Darod clan of the same name (i. e.
the Ogaden) live groups of Obo Galla 3); and with the Somali Habar
Awal clan of the BritishProtectoratewho extend into the same region
live Galla Waramiyo, Warapito (or Warabito), Warakiyo, Igo, and
Waralogo. Finally, the Rer Roble Bdrama living amongst the Somali
Gadabursi clan of the west of the British Protectorateare said to be
Galla, and there may well be other groups4).
The extent of admixture and the character of incorporationof
these Galla communitiesliving amongstSomali varies very much from
place to place. With the incorporationof Garso and Babile Galla, the
Dardd Geri of the Jigjigaregionhave split up into two separate groups,
one to the northof Jigjiga,and the otherto the south. In each the Geri
Somali are politicallydominant,but the Geri Garso, who in contrast
to the Geri Babile are largely agricultural,are Galla-speaking5).
I have directpersonalknowledgeof such residualGalla communities
only amongst the Habar Awal of the west and centre of the British
Protectorate. Here these are not Galla-speakingand are closely assimi-
lated to the Somali of the area. Indeed physically- at least to the
eye - and culturallyand linguistically,they are indistinguishablefrom
the Somali amongst whom they live and with whom they are heavily
intermarried. Amongstthe Habar Awal Sa'd Muse cultivatorsin the
Hargaisa-Gebile-Boramaregionin the west of BritishSomaliland, they
are distributedin small autonomous farmingcommunities,or resident
as individualsin the cultivatingvillagesof the local Habar Awal Somali.

" "
x) These figuresare taken fromthe Jigjiga District Notebook
with the permissionof the DistrictCommissioner, Hargaisa.
" "
2) I use the term clan-family to referto the largestSomali lineage
divisions.
3) Cf. Pirone, op. cit.,p. 141. "
4) Informationlargely taken from Jigjiga District Notebook .
, Memorialedition,
5) See R. Burton, First Footstepsin East Africa "
London,1894,vol. i, p. 192; B. Francoijni, I Somali del Harar, Africa
Italiana ", 1938, pp. 1, 127-29.

This content downloaded from 193.205.142.142 on Mon, 01 Feb 2016 09:56:47 UTC
All use subject to JSTOR Terms and Conditions
 IN NORTHERN
THE GAI^IyA SOMAI,II,AND 25

In this area the Akiso,the largestGalla group,probablynumberno more

than a fewhundredindividualsand are widelydispersed. Yet theyhave
their own headmen recognisedby the Governmentand their own inde-
pendent organizationfor paying blood-compensation. They are thus a
" "
diya-payinggroup x) in their own right, but it seems that they
sometimeshave to depend upon the support of the strongerSomali
with whom they live. They enjoy rightsto arable land which are only
weaker than those of their numericallysuperiorSomali hosts in virtue
of their numericalweakness and inabilityto meet a concertedattack
from their Somali neighbourswithout the assistance of the Adminis-
tration. But theyare not generallymolested,and usually live peacefully
amongst their hosts being identifiedwith them in external political
relations. As protected minoritiesthey are assigned the traditional
status of clients magan)., and sometimeswith other alien minoritiesare
referredto derogativelyas gun (literally" bottom") as opposed to the
" "
numericallystrong who enjoy the status of aristocrats {gob).
The cultural and social assimilationof these Akiso minorities(and
of course other, non-Galla minorities)to their strongerhosts is rarely
accompanied by genealogical identification. And here it is necessary
to explain that the Northernpastoral Somali have a segmentarylineage
organizationin which political relations are expressed partly in gene-
alogical terms,and partly througha system of political contracts (sg.
her)whichcut across genealogicalrelationships2). Minoritiescan achieve
securitythrough contract without genealogical assimilation; and this
is generallythe position with the Akiso in question when they form
wider alliances with neighbouringSomali.
In a few cases, however,such Galla clientshave been incorporated
in the genealogies of their Somali hosts. Thus with the Habar Awal
Rer Adan 'Abdalle, who are nomadic,Jjvea group of Akiso dependents
who in 1946 wereinvolvedin a fightwith anothersegmentof the Somali
Habar Awal clan. Seven of the latter were killed,but only threeof the
Akiso who were requiredto settle fouroutstandingbloodwealthsvalued

*) This is the name given in the Protectorateto those agnatic kins-

men who collectivelymeet theirblood debts (includingfull blood-wealth,
Arabicdlya,Somali mag) in concert. The diya-payinggroupis the funda-
mentaljural and politicalunit of NorthernSomali society. In 1958 in the
Protectorate therewereover 360 such groupsin a total populationestimat-
ed at about 640,000. Theirsignificance is discussedin my paper Clanship
and Contractin NorthernSomaliland," Africa July,1959,pp. 274-293.
2) See Lewis, op. cit., 1959.

This content downloaded from 193.205.142.142 on Mon, 01 Feb 2016 09:56:47 UTC
All use subject to JSTOR Terms and Conditions
26 I. M. IyEWIS

at a hundredcamels each x). This they had not the resourcesto pay,
and they approached theirAdan ?Abdalle hosts seekingassistance from
the latter and full genealogicalincorporationin their genealogy. Ulti-
mately, after much discussion and opposition from other Akiso who
wished the segmentconcernedto preserveits Akiso identity,they were
fully assimilated to the Habar Awal clan and are now regarded as a
segmentof the Adan cAbdalle lineage. This loss to Akiso integritystill
rankles with other Akiso.
Another case appears to occur with the Hinginle lineage of the
Dulbahante clan in the east of the BritishProtectorate. Accordingto
some, the Hinginle,who are an autonomousdiya-payinggroup and who
figurein the Dulbahante genealogy,are in fact of Galla Akiso origin.
But this is less well authenticatedthan the previous case.
More typically,such dispersed Akiso minoritiesretain their own
independent genealogies which, unlike those of the Somali, usually
go back only four or five generationsto their eponym Akiso. Their
short genealogies are in keeping with their small numbersand slight
political significance. For in NorthernSomaliland there is a general
correlationbetweenthe size, and thereforefightingstrength,of a group,
and its political importancea). A few Akiso, however, trace descent
"
quasi-historicallythroughmany generationsfrom Samale the foun-
der of the Somali nation3). The followinggenealogyof an Akiso elder
of the Gebile region(to the west of Hargaisa in the BritishProtectorate)
was conservedin writing(in Arabic) in a Quran.
e
Aqll Abu Talib 4)
Samale Warre
Abubakr Samale
Hagi Abubakr
Madahweine Hagi
Mihi Madahweine

x) The standardvaluationof blood-compensation fora man is a hun-

dredcamels,and fora woman,fifty.The amountsactuallypaid, however,
vary considerablyaccordingto the contextin which a homicideoccurs.
Withina dia-payinggroup,blood-wealthis usually considerablyreduced
in keeping with the solidarityof the group.
2)" See my articleForceand Fission Somali LineageStruc-
in Northern
"
, AmericanAnthropologist (forthcoming).
ture
3) See 1/EWlS,The Somali Lineage Systemand the Total Genealogy ,
(cyclostyled),Hargaisa, 1957.
4) The majorityof the Somali people trace descentthroughcAqflto
Abu Talib who died in Mecca in A. D. 620. Intervening generationsbet-
ween Sample and eAqilhave been leftout in the diagram.

This content downloaded from 193.205.142.142 on Mon, 01 Feb 2016 09:56:47 UTC
All use subject to JSTOR Terms and Conditions
 THE GAIXAIN NORTHERN
SOMAIJI^AND 27

Same Mihi
Mah Same
'Abdillahi Mah
Si'id 'Abdillahi
Ahmad Si' id
eAbdi Ahmad
Mubaraq 'Abdi
Qalib Mubaraq
Akiso Qalib - Eponytnof the Akiio
Kuttaye Akiso
Alio Kuttaye
Dado Alio
Adelle Dado
Fahiye Adelle
Sirwa* Fahiye
fAye Sirwa'
Hirsi 'Aye - Living Aki§o elder.

Those residual Galla ' communitiesliving amongst the Somali of

the west of the Protectorateand in Hargaisa District claim to have
been the originalinhabitantsof the countrybeforethe Somali. Those
who now cultivate amongst the agriculturalHabar Awal, however,
werelike the latteroriginallypastoralists,forthe cultivationof sorghum
which is now extensivelypractised in the west of the Protectorateis
scarcely older than fiftyor sixty years. Plough cultivation has been
adopted fromthe Galla of Ethiopia throughSomali in the Harar-Jigjiga
regionand has broughtvariousGalla culturetraits associated with agri-
culturein its train. Thus Somali of the west of the Protectorateuse a
plough of Galla type x) and sing ploughingand threshingsongs in which
there are many Galla words although independent Somali versions
now also exist.
At harvest time, work-partiesconsistingof kinsmen and neigh-
boursand called gus (cf. Galla guza) are summonedto threshthe sorghum
in each farmer'sfield. Such work-partiesare not foundamong the Nor-
thern pastoral Somali where most collective enterprisesare organized
on a basis of clanship (tol) and contract(her). These examples of Galla
influenceare all recent innovationsand not evidence of an earlieroc-
cupation of the area by Galla cultivators.

*) For illustrationsof Galla and Somali ploughssee Huntingford,

loc. cit., 1955, p. 72.

This content downloaded from 193.205.142.142 on Mon, 01 Feb 2016 09:56:47 UTC
All use subject to JSTOR Terms and Conditions
28 I. M. LEWIS

The real evidence for the earlierGalla occupation of these regions

is thus not the adoption of Galla culturetraits,whichis recent,but the
presenceof these residual Galla communitiesfar fromthe main centres
of the Galla in their present distributionsupplementedby the local
Somali traditionswhich regardthe Galla as their predecessors. In the
Jigjiga area itself,traditionstill maintainsthat the Galla preceded the
Somali there; and the earliest wells in Jigjiga, Marar, El Yare, and El
Bahay, are all attributedto the Akiso. Further evidence is provided
by a number of local place-names which are clearly Galla and which
are regardedby the local Somali as Galla. Such are Tog Wa gale; I gara
(perhaps fromGalla igar, to build); Taise (perhaps fromthe Galla td'
to sit or settle); Gara Horato (fromgara, people, and hor,to multiply
or prosper); and even Hargaysa, the capital of the BritishProtectorate
(cf. Galla hargaysa, a plant) x). No doubt thereare otherexampleswhich
I did not recordbut these are place-nameswhich Somali of the region
single out as of Galla origin.
This new material taken with the traditionsrecentlyreportedby
Dr. Pirone2) fromthe Ogaden, and by CerullifromMijerteinia3) provides
substantial evidence for a prior Galla occupation of western,central,
and eastern NorthernSomaliland, and is corroboratedby other local
traditions which I mention presently.

III.
FROMNORTHERNSOMALILAND.
THE DIR AND THE GALLAWITHDRAWAL

Having established that the Galla occupied NorthernSomaliland

as well as (perhapslater) centraland southernSomalia beforethe Somali,
I now seek to trace somethingof the circumstancesof their expulsion
fromNorthernSomaliland. Traditionsfromall the areas today occupied
by Somali concurin regardingthe Dir clan-familyas the oldest Somali
stock4). Traditionallytheir firststrongholdwas along the coast and
inland of what is today Erigavo Districtin the east of the BritishProtec-

of thesewords
*) Possiblya speciesof aloe. The Galla interpretations
are based mainlyon A. W. Hodson and C. H. Walker, A Grammarof
theGalla or OromoLanguage, London, 1922. Mr. B. W. Andrzejewskihas
also supplied valuable comments.
2) Pirone, op. cit., 1954, PP- 120-24.
3) Cerulli, op. cit., 1931, p. 154 f¥.
4) Cf. M. Colucci, Principi di DirittoConsuetudinariodella Somalia
Italiana Meridionale,Florence,1924, p. 99; Lewis, op. cit., i955> P- 25*

This content downloaded from 193.205.142.142 on Mon, 01 Feb 2016 09:56:47 UTC
All use subject to JSTOR Terms and Conditions
 IN NORTHERN
THE GAI/I^A SOMAUI,AND 29

torate. But in this regiontoday only a few very small Dir remnants
survive. Such are the Gambelle,Magadle, Madigan 1), Irdodub, Madobe,
Minsulk, Turre and Gurgure,of whom only a few individuals of each
groupseem to survivetoday in Erigavo Districtwherethey are attached
to, and often completelyidentifiedwith, the strongerlocal Isaq and
Darod clans with whom they live. Other strongerDir groups survive
in the Her and Gadsan, both lineages of sheikhs and holymenliving
as pastoral nomads and priests amongst the Dar5d Ogaden.
In the west of the British Protectorate,among the Habar Awal
cultivators,other small scatteredgroups of Dir survive in the Madigan
who have a similar client status to the Akiso (Galla) minoritiesof the
same area. To the east of Harar live the Bursuk, a small clan also of
Dir originbut apparentlytoday very mixed in composition,and mainly
practising cultivation2). The GurgureDir also occur in some strength
in the same region8).
But the strongest Dir clans today in Northern Somaliland are
the 'Ise (35,000 under BritishProtection)and Gadabursi (45,000 under
British Protection)in the west of the Protectorate,Harar Province of
Ethiopia, and French Somaliland. The graves of theirclan eponymslie
in Erigavo Districtin the east of the Protectorate,several hundredmiles
to the east of theirpresenthabitat. There are also various Dir commu-
nities in southernSomalia, but the largest and most importantare the
BImal clan of Merka region. The presentscattereddistributionof the
Dir raises the question of how the Dir came to vacate their original
strong-holdin the north-eastof Somaliland and how theirmovements
affectedthe earlier Galla inhabitants.
Under pressurefromSomali - at firstapparentlymainlythe Dir -
the Galla moved in two directions. In a southerlydirectionthey moved
throughcentraland southernSomalia whereat least in some areas they
encounteredpre-Hamitic Bantu and perhaps Bushmen-typepeoples,

x) Hawiye traditionrecords a group called Madagan, of allegedly

Arabian origin,as participatingin the early historyof Mogadiscio(See
CERUIAI,Somalia. ScrittiVari Editi ed Inediti,vol. ii, Rome, 1959,pp. 237-
41). Since the Dir ultimatelyclaim an Arabianoriginas do otherSomali
groups, tracingtheirdescentthrough'Aqll Abu Talib, it is possiblethat
the " Madagan" may be the Dir Madigan.
2) On the Bursuk see Burton, op. cit., i, p. 193.
3) The Gurgurewho are evidentlymuchmixedwithGalla are classified
as EasternGalla by Huntingford(op. cit., 1955, p. 15). They are mentio-
ned twice in the Futuh al-Jiaba§a (ed. and trs. R. Basset, Paris, 1897-
1909).

This content downloaded from 193.205.142.142 on Mon, 01 Feb 2016 09:56:47 UTC
All use subject to JSTOR Terms and Conditions
30 I. M. I^EWIS

whom they expelled, but of whom a few residual communitiesstill

survivetoday 1). To the north-west theymovedinto what is today Harar
Province of Ethiopia. Huntingford 2) argues that the Galla migration
fromSomalilandbegan at least as earlyas the 12thcenturyand describes
the penetrationof the Harar Highlands by what are today the Eastern
Galla, a movementwhich by the end of the 16th centuryhad taken
them across the Bashilo River and into the very heart of Ethiopia.
Contemporaryoral traditionin NorthernSomaliland knows little
of this Galla retreatand Dir advance since these eventshave been overs-
hadowed by the more recent expansion of the Darod and Isaq clan-
familiesat the expense of the Dir, a phase of Somali historyto whichI
referpresently. Traditionholds, however,that when Sheikh Aw Bark-
hadle, of Sarifdescent,arrivedin NorthernSomaliland to teach Arabic
and Islam and settledat the place whichtoday bears his name - " Aw
Barkhadle" - some thirtymilesto the north-west of Hargaisa he fought
the Galla and defeated them8). Tradition dates the] Sheikh's arrival
in the i2/i3th century. But since he seems to have been, as Cerulli
suggests4), one of the |ancestors of the Walasma* dynastyof Ifat, and
indeed accordingto an Arabic chroniclewhich I discoveredin Somali-
land 6) the fifthascendingancestorof eUmarb. Dunya-huz, the founder
of the dynastywho died in 1275 A. D., it appears that his strugglewith

x) See Ceruuj, op. cit., 1926; and Gruppietnicinegrinella Somalia,

" Archivio
per l'Antropologiae FEtnologia lxiv, 1934. See also V. L.
GroTTANEI,U, PescatoridelVOceanoIndiano, Rome, 1955, pp. 35-43; and
Lewis, op. cit., 1955, pp. 41-44.
2) Huntingford, op. cit., 1955, p. 19.
3) For a briefnotice of the Sheikh'sworks see Lewis, " Bulletinof
"
the School of Orientaland AfricanStudies, 1958, xxi/i. p. 135. Sheikh
Aw Barkhadleis traditionally the Arab proselytizerwhosubduedthe power
of the Yibir - who are now bondsmenand widelydispersed- by over-
comingtheirancestorin a sortof ritualduel. This traditionwhichconfirms
present-daycustomarypractise,explainingthe alms paid to the Yibirs
by Somali as part of the blood-priceowed to themin perpetuityfor the
murderof theirancestor,also suggeststhat the Yibir are pre-Somaliand
possiblyGalla. But I have no specificevidenceof this. On this pointsee
also Pirone, op. cit., 1954, P- 120•
4) Ceruwj, " RRAL Ser. 6, iv, 1931, p. 67.
5) This is a list of the rulersof the Land of Sa'd ad-Din appendedto
an unpublishedhistoryof Harar writtenoriginallyapparentlyby Ibrahim
Sharifof Harar. The Arabicmanuscriptto whichI had access was owned
by Sheikh'Ali SheikhIbrahimof the EducationDepartmentof the Soma-
lilandGovt and was copiedby himfromthe versionof anothersheikhtaken
fromthe original.I hope to publishthis documentat a later date.

This content downloaded from 193.205.142.142 on Mon, 01 Feb 2016 09:56:47 UTC
All use subject to JSTOR Terms and Conditions
 THE GAU,AIN NORTHERN
SOMAUI,AND 31

the Galla was probablyin the 12thcentury.1). Thus if|thisis a correct

interpretation, therewere at this time still Galla in the vicinityof Har-
gaisa whose name itselfis, as has been pointed out, probably of Galla
origin.
Further light on the Dir advance and Galla withdrawalseems
to be affordedby an Arabic manuscriptdescribingthe historyof the
Gadabursi clan2). This chronicleopens with an account of the wars
of Imam 'Ali Si'id (d. 1392) fromwhomthe Gadabursi today trace their
descent and who is describedas the only Muslimleader fightingon the
Western flank in the armies of Sa'd ad-Din (d. 1415), ruler of Zeila.
These campaigns were clearly against the ChristianAbyssinians,but it
appears fromthe chroniclethat the Gadabursi were also fightingthe
Galla. A later leader of the clan, Ugas cAli Makahil, who was born in
x575 Dobo, northof the presenttown of Borama in the west of the
British Protectorate,is recordedas having inflicteda heavy defeat on
Galla forcesat Nabadid, a village in the Protectorateon the boundary
betweenthe Habar Awal and Gadabursiclans in theirpresentdistribution.
At the same time in their oral traditions,the Gadabursi record
a gradual expansion with the fIse along the coast of the Red Sea from
the east of Somaliland,a movementin the courseof whichtheyadvanced
inland in the west expelling the earlier Galla Garso and Akiso of the
region. This traditional account of their early movements seems to
accord with the distributionof the places of birthand of burial of the
early heads of the Gadabursi clan as relatedin the Gadabursi chronicle.
For, as has been pointed out, the eponymof the Gadabursi is with the
eponymof the 'Ise clan buriedin the north-eastof the Protectoratenot
farfromSheikhIsaq's tomb in Erigavo District; and later leaders of the
Gadabursi are buried generallyto the north of Borama, their present
headquarters in British Somaliland.
I have referredto a parallel movementof the *Ise Dir fromeast
to west. *Ise traditionstell of conflictbetweentheirancestorsand those
of the Isaq clan-familyin the east of the Protectorateas a result of
which the *Ise moved westwards towards their present position. If

x) See Ceruuj, op. cit., 1931,p. 40. On theWalasma*dynastysee also

J. S. Trimingham,Islam in Ethiopia, Oxford,1952, pp. 58-60.
2) This is a shorthistoryof the leadersof the Gadabursiclan written
by SheikhAbdarahmanSheikhNur of the GadabursiRer Ugas and based
on an earliermanuscriptby Sheikh BarkhadleSheikh 'Abdillahi,also of
the Rer Ugas. Sheikh Abdarahmanis the inventorof the " Gadabfirsi
"
script",see myarticleThe GadabuursiSomaliScript , Bulletinofthe School
of Orientaland AfricanStudies", 1958,xxi/i, pp. 134-56.

This content downloaded from 193.205.142.142 on Mon, 01 Feb 2016 09:56:47 UTC
All use subject to JSTOR Terms and Conditions
32 I. M. IvEWIS

these traditions are historicallyvalid they imply that the *Ise only
expanded westwardsat a fairlylate date, subsequent to Sheikh Isaq's
reputed arrival at Mait in the east in the 12/13th centuryto which I
referbelow.

IV.
OF THE DIR FROMTHE EAST AND THE EXPANSION
THE MIGRATION
OF THE DARODAND ISAQ ClyAN-FAMIUES
.

The dispersal of the Dir clans to the west and to the south under
pressurefromthe expanding Dar5d and Isaq communities,a phase of
Somali historywhichto some extentat least coincideswith the rout of
the Galla and theirmigrationsinto Ethiopia, bringsus to a more recent
stage in Somali historybut still one for which firmdates are lacking.
That Darod the ancestorof the clan-familyof the same name, preceded
Sheikh Isaq, founderof the Isaq clans, is universallyheld by Somali.
The precedenceof Dar5d is also reflectedin the commonphrase " Dir
"
iyo Darod (Dir and Dardd) used still to distinguishtwo great branches
of the Somali nation.
Traditionally,the ancestor of the present Darod clan-family,a
group over a millionstrongand widely distributed,was SheikhDarod
Ismail, son of Ismail Gabarti,a descendentof Abu Talib through'Aqil,
who is said to have crossedto the north-eastern Somali coast fromAra-
bia about the 10/nth century1). The exact place of his arrival varies
in the traditions,but accordingto one generallyheld account, Darod
firstsettled at a place called " Darod Ful " near Bosaso. Darod made
contact with the Dir who were then still in this regionand marrieda
Dir woman called Donbirra fromwhose sons the present Darod clans
and lineages stem 2). Tradition recordsthe growthin strengthof the
Dar5d at the expenseof the Dir, and the latter'sdispersalfromthe north-
east cornerof Somaliland.

x) There are various hagiologiesin circulationin Somaliland,some

printedin Arabic, which identifyDarod Ismail with the son of Isma'Il
Gabarti,the well-knownArabiansaint who died in 1403 and is buriedat
Zabid in the Yemen (e. g. the Mandqib as-Sayh Isma'tl bin Ibrahimal-
Gabartt,Mustafa al-Babl al-Halabi Press, Cairo, 1945, writtenby Sayh
Ahmad bin Husen bin Mahammad). This seemsclearlyan impossiblecon-
withthe local traditionsproclaiming
nectionsince it conflicts the antiquity
of the Darod and would make the Darod later than the Isaq.
2) On thelegendsconcerning Darod's receptionby the local Dir Somali,
see I/EWIS,op. cit., 1955, PP- 18-19.

This content downloaded from 193.205.142.142 on Mon, 01 Feb 2016 09:56:47 UTC
All use subject to JSTOR Terms and Conditions
 THE GAUAin NORTHERN
somai,h,and 33

As Cerulli has argued, some idea of the early distributionof the

Dardd and of the directionfollowedin their expansion can probably
be gained fromthe distributionof the graves of Darod clan ancestors1).
The grave of Dardd himself,a roughstone structureand not an elaborate
domed tomb such as that of Sheikh Isaq, lies in extremelydesolate
countrysome twelve miles to the south-eastof the villageofHadaftimo
in Erigavo District of the BritishProtectorate,and some two miles off
the Hadaftimo-Buranroad. Harti, Darod's grandson,and founderof
the great Harti division of the Darod clan-familyis buried in a grave
of similarconstructionbut of smaller dimensionsa furthertwelve miles
south-south-east.Accordingto Cerulli,he perishedat the hands of the
Galla who were contestinghis power in the area. My informantsdid
not knowthis tradition,but whenI visitedthe gravepointedto a mound
of stones near Harti's grave in which I was told his arm was buried.
He has lost his arm,theysaid, in a gihadwiththe people of the area who
were not specified (but may have been Galla) and perished shortly
afterwards.
The distributionof the graves of subsequent Darod ancestors,
the descendantsof Harti, suggestsas Cerullihas pointed out, a gradual
movementof the Dar5d to the south-east. This expansion inland must
have been underwayabout the n/i2th centuryif the traditionaldate
of Darod's arrival is correct2); and accordingto traditionscollectedby
Cerulli in Mijerteiniawas accompanied by the expulsion of the earlier
Galla inhabitants,and (frommy own evidence) also of the Dir Somali.
The great movementsof the Darod into central and southernSomalia
have been charted by Cerulli,but their expansion in northernSomali-
land has hardlybeen touched upon. It is this northernmigrationwhich
concerns us here.
The firstsecure mentionof the Darod is in Sihab ad-Din's 16th
centuryFutiih al-HabaSa9) which constantlyrefersnot to the Darod
as such, but to that branchof them descendedfromHarti, and to other

" AfricaItaliana
x) Ceruuj, 1931, p. 158.
2) I have suggestedabove that the connectionbetweenthe eponym
of the Darod clan and Isma'il GabartIof Zabid is mythical.This does not,
however,necessarilyinvalidatetheirpretensionsto Arabian ancestry. It
is possiblethat theirancestorwas of less renownthan the Zabid saint,and
that they should seek to attach themselvesto the latterforthe sake of
prestige. And even if the traditionsofdescentfroman immigrant Arabian
are entirelyfalse,it seems to me that the traditionsthemselvesmust be
construedas representing an expansionof the Darod at about this time.
8) Ed. and trs. R. Basset, Paris, 1897-1909.
3

This content downloaded from 193.205.142.142 on Mon, 01 Feb 2016 09:56:47 UTC
All use subject to JSTOR Terms and Conditions
34 I. M. IyEWIS
"
individualDarod clans. In the Futuh, the Harti are describedas the
people of Mait ", Mait being a small sea-port and today ruined city
(probablyof great antiquity)x) on the coast of Erigavo in the British
Protectorate. Of the most westerlyof the Darod clans, those now living
in the Harar-Jigjigaregion,the Geri (who are traditionallysupposed
to have arrivedtherein the 16th centuryin the train of Imam Ahmad's
victoriousarmies),the Bartirre,and Yabarre 2) are mentionedby name
in the Futuh. Their whereaboutsthen is not specified,nor is it clear
fromthe text, but they must have been in contact with the Marrehan
Darod who today live south of the Ogaden, since a dispute between
the Geriand Marrehanis mentionedin the Futuh. This suggestsI think,
that the Marrehan were then north of their present position - per-
haps in the Ogaden region- and expandingin the centre of Northern
Somaliland while the Geri were movingfurtherto the west with other
Darod groups.
Traditions describingthe occupation of the Ogaden region itself
by the Darod clan of that name (the Ogaden) have recentlybeen pub-
lished by Pirone3). From their originalnorth-easternseat, where ac-
cordingto my informants theirancestoris buriedto the east of Erigavo,
the Ogaden had moved southwardsinto what is now the Mudug Provin-
ce of Somalia, and fromthere had pressed north-westwards along the
vallies of the Fafan and Gererto the westernextremitiesof the Ogaden.
In this expansionthey had to strugglewith the Raitu and Arussi Galla
and with groups of Dir and Hawiye 4) for possession of pasturage and
wells. It would thus appear that the Darod must have occupied the
Ogaden by the 17th century,and probably have been there for some
time, since it was only towardsthe end of the next centurythat they
were able to cross the Shabelle River against strongHawiye resistance
and move into the Doi region of southernSomalia5). In any event,
the Dir who by the 17th centuryand probablyearlierhad been driven
out of most of north-eastern and centralSomaliland, left behind them,

1) As well as figuringin the Futuh,Mait is markedon some 17thcen-

turymaps of the Horn of Africa(e. g. in the " Blaeu Atlas Amsterdam,
1665).
2) This seems a more likelyreadingforthe Arabic than Bas-
set's " Yibberi
3) Pirone, op. cit., 1954, P- I22-
4) As far as I knowthe only Hawiye grouptoday presentin the east
of NorthernSomalilandis the Hawiye Rer Fiqi Sinni who live amongst
the Dulbahante of the Protectorate.
5) See Ceruwj, op. cit., 1926.

This content downloaded from 193.205.142.142 on Mon, 01 Feb 2016 09:56:47 UTC
All use subject to JSTOR Terms and Conditions
 IN NORTHERN
THE GAIylyA SOMAUI.AND 35

as the Galla had also done, indelibletraces of theirprevious occupation

in the residual Dir groups which are today found among the Ogaden.
But if local traditionis correct,despite the development of the
Darod during the nth and 12th centuries,the Dir cannot have been
decisivelyroutedin what is today Erigavo Districtof the Protectorate.
For when Sheikh Isaq, the founderof the presentIsaq clans, settled in
the ancient city of Mait in traditionallythe i2/i3th century,the area
surroundingthe port was still in the hands of the Dir, althoughno Galla
are rememberedas having still been in the vicinity*). Moreover,Isaq
married a Dir woman of the Magadle clan (see above p. 29) whose
progenyformone of the great divisionsof the Isaq clans - the Habar
Magadle 2). It would thus seem that the Darod were then still confined
furtherto the east in the northof what is now the MijerteinProvince
of Somalia.
The ruins whichlie some little distance to the west of the present
port of Mait and close to Sheikh Isaq's strikingtomb whichis a famous
place of pilgrimage,are traditionallythe remains of the settlement
foundedby Sheikh Isaq and his followersand built probablyupon ear-
lier foundations8). Tradition states that when Sheikh Isaq arrived at
Mait fromthe west wherehe had spent many years aftercrossingfrom
Arabia to Zeila4), he foundthe Magadle Dir in possessionof Mait and

*) Some historicalconnectionbetweenthe Darod and Isaq can be

establishedby the fact that the Dulbahanteand fourotherminorDarod
groupstracedescentfromthe unionof theircollectiveancestor,Harti Kom-
be, witha daughterof Arab, a son of SheikhIsaq. Accordingto present
Darod genealogiesHarti Kombe is the son of Darod's grandson,Kombe
Kablallah. And this suggestsa gap of at least about one centurybetween
SheikhIsaq and SheikhDarod.
2) The Habar Magadleare the onlyIsaq mentionedin the Futuh
" group "
al-gabaSa. At least it seems that the Habr Maqadi (<3^** des-
cribedin the Futuhare the Habar MagadleIsaq. The Isaq are notmentioned
underthe name Isaq, nor is thereany reference to the easterlybranchof
the Isaq, the Habar Habused.
8) At Mait fragments of Chineseporcelainare to be foundscattered
about on the topsoilroundthe ruins. These urgentlyrequireexcavation,
and sincethe portis referred to in the Futuhits excavationand the study
ofits connectionswiththeotherruinedtownsofNorthern Somalilandshould
throwveryimportantlighton a little-known aspectof Somali history.For
a discussionof a preliminary investigationof some of the moreimportant
NorthernSomali ruins,see A. T. Curiae, The RuinedTownsof Somaliland ,
"
Antiquity", Sept., 1937, PP- 315-*7-
4) A very circumstantial account of the movementsof bheikhIsaq s
forefathersin Arabia and of the Sheikh'sown peregrinationsbothin Arabia

This content downloaded from 193.205.142.142 on Mon, 01 Feb 2016 09:56:47 UTC
All use subject to JSTOR Terms and Conditions
36 I. M. I/EWIS

rulingthe other Dir groups whose present-dayrepresentativesin this

regionhave been mentionedabove (p. 29). And as the Isaq expanded
these earlier Dir communitieswere drivenwestwardsand to the south
towardstheirpresentpositions,just as previouslythey had moved from
the northeastcorner of Somaliland under pressurefrom the growing
Darod community. In their train the Isaq spread westwardsand to
the south-west,but did not penetrateas far in the latter directionas
the Darod. In this generalexpansionthe Isaq split up into theirpresent
componentsegments,one fractionof the Habar Yunis clan, the Habar
Yunis Muse 'Arre,remainingbehindin the Mait regionof Erigavo Dis-
trictas the custodiansof the tomb of theirfounder,Sheikh Isaq. But
it seems that if the Isaq were in fact in the Mait regionin the 13th cen-
tury as their traditionsmaintain,they must later have for a time lost
Mait to the Harti Darod since it is they who are describedin the 16th
"
centuryFutuh as the people of Mait

V.
CONCLUSIONS.

The material presentedin this paper makes it fairlycertain that

the people whom we now know as Galla did in fact occupy much of
NorthernSomaliland beforethe Somali. But by the time of Sheikh
Darod's traditionalarrival fromArabia in the 10/nth century,(or the
growthof the Darod communityabout this time) the north-eastern co-
astal strip was occupied not by the Galla but by the Dir Somali who

priorto his arrivalin Somalilandand in Ethiopiaand Somalilandsurvives

in local tradition.Thesetraditionsare recordedin Arabicin severalmanus-
criptand publisedhagiologies. One of the mostrecentof the latteris the
amgad of Sayh Husen bin Ahmad Darwis al-Isaqi as-Somali,Aden, 1955*
For a life of the saint by one who has spent many years collectingthese
traditions,see cAu Sheikh Mahammad,The Originof the Isaq peoples ,
" The Somaliland
Journal", Vol. 1, 1954, PP- 22-6 (Hargaisa, BritishSo-
maliland). So far thesetraditionshave not been corroborated by any ex-
ternalevidence,and in the opinionof myfriendProf.R. B. Serjeantof the
School of Orientaland AfricanStudies,London,would not be acceptedas
historicallyvalid by Arabian historians. Thus, as in the case of Sheikh
Darod, it seemsunlikelythat the historicaloriginsclaimedforSheikhIsaq
are authentic.Nevertheless, I believethattheyhave an important historical
contentat least in the sense that theyrepresentan earlyexpansionof the
Isaq groupand bringthisinto some sortof historicalrelationship withthe
earlierextensionof the Darod.

This content downloaded from 193.205.142.142 on Mon, 01 Feb 2016 09:56:47 UTC
All use subject to JSTOR Terms and Conditions
 THE GAU,AIN NORTHERN
SOMAIJI,AND 37

still lived in the Mait regionfurtherto the west whenSheikhIsaq settled

there in the i2/i3th century. The Galla would seem thus to have al-
ready begun their westerlymigrationsat about this time as Hunting-
fordsuggests: certainlythey were no longerdominantalong the north-
eastern coast of NorthernSomaliland. But they were still apparently
to be foundin the Hargaisa regionin the 12thcentury. And inland they
must also have still been in the east if the Dardd traditionsof conflict
with them are valid. They would appear, however,to have been driven
furtherwestward not long after. And later, the majority of the Dir
were also expelled fromnorth-eastern Somaliland,probablyin the 13th
and 14th centuries.
By the 17th centuryat least, the Darod and Isaq seem to have
taken over much of north-eastern and centralNorthernSomaliland and
also the Ogaden region. Other Darod had by this time penetratedas
far west as Harar, where there were already Dir Somali established;
and in the Jigjiga regionthe Galla who had earlierdominatedthe area
were outnumberedby Somali immigrants.
Thus the present distributionof the NorthernSomali seems to
have been assumedby the i6/i7th century,as is to be expectedconsider-
ably earlier than in the south where many of the Somali movements
of expansion are much more recent. It is of interestthat I found no
record in BritishSomaliland today of any of the other Somali groups
who traditionallycame fromNorthernSomaliland to settlein the south,
except in the case of the Hawiye of whoma small group- the Rer Fiqi
Sinni - are found amongst the Dulbahante (Dardd) clan of the east
of the Protectorate.
The southwardsexpansion of the Somali fromthe shores of the
Gulf of Aden still continuesdespite the establishmentof international
frontiersand Administrativecontrol. It is very evidentin the Northern
Provinceof Kenya, and in the BritishProtectoratethe Isaq now appear
to be pushingoutside the territoryat the expense of the Darod into the
Ogaden and Haud.
The briefchroniclepresentedhere of some of the more important
population movementsin NorthnerSomaliland must remain tentative
in the extreme. It is still unfortunately impossibleto establishbeyond
doubt the validity of even the few key dates referredto. And there
are also other very considerable difficulties which beset any attempt
to disentanglethe historyof the Galla and Somali. Even the precise
ethnicrelationshipof the two peoples remainsobscure. It is sometimes
held, forinstance,that the Somali are merelypeople of Galla originwho
have been Arabized by the incorporationof Arab immigrantsand by

This content downloaded from 193.205.142.142 on Mon, 01 Feb 2016 09:56:47 UTC
All use subject to JSTOR Terms and Conditions
38 I. M. IyEWTS

long and fairlyintensecontact with Arabian Islam *). And it is certain

that there has in fact been Arabian influenceof this kind. The difficul-
ties are not eased by the fact that the names " Galla " and " Somali "
are firstrecordedonly in the late 16th and early 15th centuriesrespec-
tively; and it may be quite illegitimateto project them into the past
with the assumption that they had the same ethnic and cultural con-
notations which they have today.
The truthis that at presentthereseems no means of tellingwhether
the Somali of today are as a whole derived fromthose whom we now
know as Galla, or whetherthey representdifferent branchesof a com-
mon stock which,over a long period,has become increasinglydifferen-
tiated. Some Galla in contact with Somali claim to be Dir in origin,
statingthat the Galla are Dir; and the Galla and Dir are certainlysome-
times confusedin local Somali traditions2). But whethersuch claims
are historicalor merelyunderstandableattemptsby the Galla to equate
themselveswith the Somali 8) forreasons of prestigeappears impossible
to determineat the moment. Certainlyexisting Galla groups do not,
as far as I know, figurein Dir genealogiestoday.

x) This is the viewof Burton,op. cit.,and of R. E. Drake-Brockman,

BritishSomaliland , London, 1912, p. 68; and of others.
a) Cf. PlRONE,op. cit., 1954, P- 119-
8) The Midganbondsmenwho are mainlyleather-workers and barbers
and generallydespisedby Somali, sometimesproducegenealogiestracing
descentfromDir Somali. These are regardedas fictionsby noble Somali.
Cf. K. Iy. G. GoiyDSMiTH and I. M. Lewis, A Preliminary of
Investigation
"
the Blood Groupsof the sab Bondsmenof NorthernSomaliland , Man ",
December 1958, pp. 188-190.

This content downloaded from 193.205.142.142 on Mon, 01 Feb 2016 09:56:47 UTC
All use subject to JSTOR Terms and Conditions