Acceptability and Use of

Cereal-Based Foods in Refugee Camps:
Case-Studies from Nepal, Ethiopia, and Tanzania

Catherine Mears with Helen Young

An Oxfam Working Paper
©OxfamGB 1998

First published by Oxfam GB in 1998

ISBN 0 85598 402 3

A catalogue record for this publication is available from the British Library.

All rights reserved. Reproduction, copy, transmission, or translation of any part of this
publication may be made only under the following conditions:

• with the prior written permission of the publisher; or
• with a licence from the Copyright Licensing Agency Ltd., 90 Tottenham Court Road,
London Wl P 9HE, UK, or from another national licensing agency; or
• for quotation in a review of the work; or
• under the terms set out below.

This publication is copyright, but may be reproduced by any method without fee for teaching purposes,
but not for resale. Formal permission is required for all such uses, but normally will be granted immediately.
For copying in any other circumstances, or for re-use in other publications, or for translation or adaptation,
prior written permission must be obtained from the publisher, and a fee may be payable.

Available from the following agents:
for Canada and the USA: Humanities Press International, 165 First Avenue, Atlantic Highlands,
New Jersey NJ 07716-1289, USA; tel. 732 872 1441; fax 732 872 0717;
for Southern Africa: David Philip Publishers, PO Box 23408, Claremont, Cape Town 7735,
South Africa; tel. 021 644136; fax 021 643358;
for Australia: Bush Books, PO Box 1370, Gosford South, NSW 2250, Australia; tel. 043 23274;
fax 029 212248.

For the rest of the world, contact Oxfam Publishing, 274 Banbury Road, Oxford OX2 7DZ, UK.

Published by Oxfam GB, 274 Banbury Road, Oxford OX2 7DZ, UK

Oxfam GB is registered as a charity, no. 202918, and is a member of Oxfam International.

Typeset by Oxfam CSU
Printed by Oxfam Print Unit

This book converted to digital file in 2010

Acknowledgements 4 3 Case-study: Somali refugees
Preface 5 in Ethiopia (May/June 1997) 66
The context 66
Executive summary 8 The field study 74
Abbreviations 10 Conclusion 88
3(a) Ethiopia: regions and zones 90
1 Main report 11 3(b) Location of camps, Somali Region 91
Introduction 11 3(c) Characteristics of the refugee
Methodology 12 population 92
Cereal fortification and blended food 13 3(d) Refugee numbers 93
Food and nutrition in the study sites 14 3(e) Market prices at time of study 94
The 'pragmatics of preference' 18 3(f) Chronology of malnutrition and
The use of food and micronutrients 20 micronutrient deficiencies 95
What made food 'acceptable'? 24 3(g) List of key informants
Cereal fortification at the regional, and interviewees 96
camp, and household levels 27
Conclusions 30
Appendices: 4 Case-study: Burundian refugees
l(a) Methods 33 in Tanzania (August/September 1997) 97
l(b) Fortified pre-cooked blended food: The context 97
definitions and examples 37 The field study 103
l(c) Prices, nutritional value, and unit cost Conclusion 118
of WFP-supplied commodities,
January 1997 39 Appendices:
l(d) Composition and nutritional analysis 4(a) Map: Kigoma Region and
of planned rations 40 location of camps 119
4(b) Age and sex composition of
population, Muyovosi camp 120
4(c) Refugees Statistics Report on
2 Case-study: Bhutanese refugees Registration, Kigoma Region,
in Nepal (January/February 1997) 41 July 1997 121
The context 41 4(d) Market prices at time of study 122
The field study 46 4(e) WFP report of market prices to
Conclusion 60 July 1997 123
Appendices: 4(f) Summary of food ration scales and kilo
2(a) Map of Jhapa and Morang Districts 61 calories, January-August 1997 124
2(b) Age and gender composition of camp 4(g) Proposed milling set-up, Isaka 125
population 62 4(h) List of key informants and
2(c) Market prices at time of study 63 interviewees 126
2(d) The process of parboiling rice 64
2(e) List of key informants and
interviewees 65 Glossary 127
References 131

This work was carried out with the aid of grants Dr Johan Pottier for advice and comments on
from the United Nations High Commissioner each case-study.
for Refugees (UNHCR) and from the Micro- Thanks and appreciation are due also to all
nutrient Initiative, an international secretariat the people from UN, NGO, and government
housed in the International Development agencies in the three field-study sites who were
Centre, Ottawa, Canada. so helpful and supportive during the field
We would like to thank Rita Bhatia, Arnold research, in particular the staff of Oxfam
Timmer, Gloria Sagarra and Janak Upadhay Nepal, Tanzania, and Ethiopia, and UNHCR
(UNHCR, Geneva) and Peter Dijkhuizen (WFP, and WFP staff in all three sites. These and oth-
Rome) for information and facilitation during ers are too numerous to mention, but many are
the course of the project. included in the list of key informants at the end
We also acknowledge the help of Dr Sue of each case study. Thanks also to the three
Chowdhury, Dr Steve Collins, and Maurice interpreters, Purnima Sharma, Ebla Abdi, and
Herson, who commented on the draft, and Mary Ruheta, for their work and friendship.

Episodes of scurvy, pellagra, and beriberi strategies they are not directly comparable. It is
among refugees during the 1980s were a stark at best awkward to discuss which strategy is more
reminder of the inadequacies and failures of the likely to succeed, because for either approach this
international humanitarian response. There depends on a different set of questions and
was a public outcry when the range and extent answers. For example, the feasibility of cereal
of micronutrient-deficiency diseases became fortification is, in the first instance, dependent on
more widely known and publicised, both in the type of cereal (rice being particularly difficult
expert meetings (Oxford, 1988) and in tele- to fortify) and the availability of local milling/
vision documentaries (such as Killed by Kindness, fortification capacity. At worst, direct comparisons
LWT, 1990). The reactions were swift, as inter- of these two optionsriskruling out completely one
national agencies and organisations sought or other option. It is far better to identify the
solutions and strategies to overcome both the limiting factors or practical constraints associated
endemic micronutrient-deficiency diseases with either strategy.
(vitamin A deficiency, iron-deficiency anaemia, Our knowledge of what refugees think and
and iodine-deficiency disorders), and the inter- do in relation to food is weak indeed. Much of
mittent outbreaks of more unusual deficiency what we know about food use and acceptability
diseases that were thought to have been all but is to a large extent based on 'received wisdom',
eradicated. There is not a single solution to the built upon over-repeated anecdotes whose
problem of micronutrient deficiencies; rather a origin can rarely if ever be traced. This received
range of strategies must be devised and adapted wisdom cannot be questioned or challenged
according to the local context.1 without proper evidence. There have been
This research represents one part of the attempts in the past to look at use and
process of developing more effective strategies acceptability of newly introduced foods,- but
to combat micronutrient deficiencies among none to our knowledge has involved in-depth
vulnerable refugee populations. The two partic- studies of emergency-affected populations,
ular strategies of interest are the inclusion of taking into account the wider determinants of
fortified blended food in die refugee food rations, use and acceptability, as this study has done.
and the fortification of cereals at local or house- A further distinguishing characteristic of this
hold level. In broadest terms, die aim of the study is the structures upon which the research
research is to provide a clearer understanding of process was based. It represents an unusual and
the reality of refugees' lives and the way in which probably unique collaboration between an
they make use of the food assistance they receive. international non-government organisation
Without such a perspective, how can outsiders (Oxfam GB); the United Nations High Com-
have any confidence that their particular missioner for Refugees; the Micronutrient
technical solutions will work? Initiative; and the World Food Programme. Such
Various approaches to solving the problems collaboration represents far more than simply a
of micronutrient deficiencies have their own secure funding base (for which we are very
shortcomings and weaknesses. For blended grateful to UNHCR and the Micronutrient
food, a problem frequently mentioned is one of Initiative). In these organisations there was a
acceptability: it is widely assumed that for most range of individuals contributing their
people it is a new food, not previously encoun- expertise and assistance, each of whom has an
tered or used. In contrast, the strategy of forti- inestimable knowledge and experience of the
fying cereals with micronutrients is far more a refugee situations under investigation. The field
technical issue, concerned with the operational researcher herself was familiar with refugee
feasibility of local fortification. Thus the practical emergencies, having worked for more than 12
problems of adopting either strategy are quite years in refugee health and nutrition, with
different, which means that as policy options or Oxfam and Red Cross inter alia. Qualitative
Acceptability and use of cereal-based foods in refugee camps

investigations of this nature require consider- were made simultaneously. For example, in
able sensitivity towards both the system of humani- Ethiopia the positive impact of adding blended
tarian response and the refugees themselves. food to the ration was negated by the simultan-
The most recent refugee crises have given eous reduction in the amount of wheat grain
rise to a climate of criticism and questioning and fortified oil. However, in Ethiopia the
about the role and effectiveness of the humani- major concern was scurvy, which was addressed
tarian response. In relation to nutrition, the focus partly by the addition of blended food (although
is often on the frequently reported nutritional the precise amounts after storage and prepar-
deficiencies, and the failures of the internation- ation are open to question). Thus blended food
al response (which are indeed described in this does indeed raise the levels of micronutrients in
report). But such a view misrepresents the the ration, but in the amounts given in these
efforts and successes of a wide range of agency particular contexts they were far from sufficient
interventions aimed at providing food assistance. to meet requirements.
Over the years, as the challenges of meeting the It is likely that the micronutrient deficiencies
needs of emergency-affected populations hasve that are reported are the tip of the iceberg,
grown almost exponentially (as have the beneath which lie the untold human costs of
numbers of people affected), many aspects of living on food rations that are nutritionally
humanitarian programmes have improved. marginal for months, if not years, on end. A
This might not be immediately apparent from a tribute must be paid to the refugees who must
reading of this document, but we should bear in daily survive on these meagre rations. Once
mind the situation ten to fifteen years ago in they receive the food, there is almost certainly
many refugee crises, such as Eastern Sudan in very little wastage, unlike the 'leakages and
the mid-1980s, where the warehouses were losses' that occur while the food is en route to
crammed with inappropriate foods, such as them. With few resources at their disposal,
slimming biscuits, special drinks, out-of-date refugee women must find ways to maintain the
tinned food, snack foods and other worthless quality of the food they receive, and maximise
items. Today such foods are the exception, not its palatability and acceptability to their families.
the norm. In their place we have a commitment At the same time they must make the ration last
by UNHCR and WFP to provide fortified foods, until the next distribution, while deciding how
such as oil, salt and blended food, which are now much to trade or sell in order to purchase other
a regular part of refugee rations, as this report essential items. Managing such a tight, often
shows. This represents a significant improve- inadequate budget is no small achievement.
ment in the nutritional quality of the rations Gaps in the food-aid pipeline and subsequent
received and is but one of many advances. shortfalls in food provision to refugees were a
Although a critique of the nutritional com- key factor affecting food consumption and use.
position of the rations encountered in these All the careful fine-tuning of rations is going to
studies goes beyond the objectives of this report, make little difference if refugees do not then
there are a few points worth noting here, receive what was promised. The reasons for the
particularly in relation to the difference made shortfalls in food provision represent huge
by the inclusion of fortified foods. In all three obstacles to any organisation. They include
sites, the salt and oil included in the ration were problems of restricted access, insecurity, and
fortified (with iodine and vitamin A respectively), most significantly a lack of resources. All exter-
which for iodine meant that the rations met the nal resources emanate from donors; in the
WHO-recommended nutritional requirements absence of clear political commitment and
for emergencies; but levels of vitamin A in the action, the technical solutions will continue to
ration were still only approaching 50 per cent address only partly the suffering of millions of
(even where blended food was distributed). The refugees affected by micronutrient deficiencies.
addition of blended food did not necessarily
result in a marked improvement in the micro- Helen Young
nutrient content of the ration, as other changes Oxfam GB

1 This is not particularly surprising; even to be extremely adaptable, being consumed
where there is an excellent means of prev- with a variety of foods both sweet and savoury,
ention, such as distribution of vitamin A and was not considered alien to their usual
supplements, this does not diminish the need foodstuffs. In contrast to the other foods pro-
for increasing vitamin A consumption through vided in the ration to these refugees, the instant
other means, such as fortification of oil, and blended food was not sold. These findings are
promotion of home gardens producing green not widely known, because the report remains
leafy vegetables. unpublished (J. Gladwin, 'Results of a Field
2 Jean Gladwin undertook field trials of an Trial of Instant Blended Food Produced by the
instant blended food among Burmese Milk Marketing Board, UK, Used in Supple-
refugees in Bangladesh in 1992/93. Her mentary Feeding Programmes for Moderately
findings concur with those of this study, as she Malnourished Children in Rohingya Refugee
found the product was very well accepted by Camps, Cox's Bazaar, Bangladesh', Liverpool
all age groups, and that the product appeared School of Tropical Medicine, UK, 1993).
Executive summary

Micronutrient deficiencies constitute a major of food, the intra-household processing and
public-health problem in refugee populations, allocation, and levels of acceptability were investi-
and fortification of the cereal staple and/or the gated, with the focus on the cereal staple and
provision of a fortified blended food are key blended food supplied in the general ration.
strategies identified for prevention. This study In all sites a proportion of food was sold by
aimed to contribute to current discourse on refugees, the amount judged in two cases to be
these prevention strategies by providing field- within 'acceptable' limits and not adversely
based data about use and acceptability of cereal- affecting the local economy. In the third case,
based foods (cereals and fortified blended food), the camps had become integral to a complex
which is crucial to successful implementation. regional and political economy.
The study also reviewed opportunities for the Blended food was sold on a very small scale
fortification of cereals with micronutrients at relative to the sale of other commodities. When
local or household level. Field studies were told of the sales, women expressed surprise and
undertaken in three sites (7-8 weeks in each): disapproval of those who were allegedly selling.
• Bhutanese refugees in Nepal: all camps The items purchased were mostly food items,
principally vegetables and more of the same or
• Somali refugees in Ethiopia: Kebribeyah
an alternative staple. There was a range of
and Hartisheikh A
reciprocal arrangements in all sites. Foods were
• Burundi refugees in Tanzania: Muyovosi
acquired for palatability and flavouring, for an
increase in variety to break the monotony, and
The visits included key-informant interviews, for special occasions.
formal and informal observation, household- Storage of meal and blended food inside the
level semi-structured interviews, briefing and home was a problem in all sites, because of the
de-briefing sessions, and reviews of secondary presence of rodents and the problem of con-
sources. tamination with dust. No preparation methods
A history of malnutrition and micronutrient were observed which seemed to be more than
disorders was compiled, together with an analysis usually detrimental to the micronutrient content of
of the micronutrient content of the planned the ration food, except multiple washing of rice
ration and comment on why the planned ration (Nepal) and sieving of blended food (only by a
was not always delivered in full and consistently. small minority). In the African sites, methods of
In all sites the planned ration was deficient in preparing whole grain and reasons for using
micronutrients, but the scale of problems of the machine mill and/or milling at home were
supply differed. There was a history of acute explained. Methods of cooking blended food
malnutrition and micronutrient deficiency in all varied between sites, but were considered
sites; incidences were not always co-existent, appropriate. Only in Nepal had refugees been
nor were they confined to one particular phase explicitly informed that blended food was pre-
of the 'emergency' programme. The quality of cooked and that cooking time could be
the monitoring of micronutrient-deficiency shortened accordingly.
disorders also differed between sites. Fuel for cooking was an issue in all cases. In
Nepal and Tanzania, where kerosene stoves
and 'improved' stoves (respectively) had been
The use and acce of introduced, women mentioned benefits related
cereal-based foo to safety and easier tending.
A preferential bias in food allocation to men
The concept of refugee preferences within the at mealtimes was implied by women interviewees.
range of ration and non-ration foods available These patterns varied according to type of
was examined. The sale, purchase, and exchange household and to circumstances peculiar to

Executive summary

daily life as a refugee. Blended food was used medium- to long-term commitment by donors;
and valued in all sites as a complementary food and considerable technical and management
for infants. All members of the household in all expertise at the milling site. Regional-level
sites were said to be eating blended food, but fortification would be most appropriate where
children tended to be given another serving the flour (milled cereal) being distributed was
later in the day if they asked. This is not the highly acceptable and mostly eaten, not sold or
same as saying that it was mostly eaten by exchanged.
children. Camp-level fortification would require ade-
Acceptability was associated strongly with quate, affordable, and accessible milling
familiarity, but unfamiliarity did not always capacity; training and health education; and
indicate low acceptability. Blended food was close monitoring of mill technicians to ensure in
highly acceptable. Even though the productp6'r particular the even mixing of the fortification
se was not familiar, it was easily recognisable as a mix. Camp-level fortification would be most
porridge and/or fashioned into an already appropriate where the staple grain being
familiar dish.The perceived and actual quality distributed was eaten in milled form and highly
of the blended food and/or the cereal staple was acceptable; or, if not, was sold or exchanged to
an issue in all sites. In Nepal acceptability of a obtain an alternative whole grain eaten in
particular ration food increased when associated milled form.
with a visible improvement in health. In all sites, Fortification of cereal at the time of grinding
blended food was identified with a nutritious or pounding at household level did not appear
porridge which they had known before. to be feasible. This was because of lack of time,
Refugees' traditional practices did not containers, space, and standard measures.
detract from the acceptability of blended food in Furthermore, the staple grain was not always
the three sites to any major extent. Differences milled to flour prior to consumption; or, if so,
in customary practices were evident between was taken to the machine mill. However, a
the educated and non-educated (Tanzania), household-level fortification mix in the form of
between pastoralists and town dwellers (Ethiopia), a salt could be added during cooking, as salt
and between the public and the private domains and/or spices were used routinely in all sites.
(Nepal). Resale value was a factor, particularly Use and acceptability of such an additive would
in the level of acceptance of the staple grain, but need to be researched.
it was not the over-riding consideration at all Locally and regionally produced fortified
times. The resale value of blended food was blended food had been distributed in the general
relatively low, while its general level of accept- ration in all sites. In Nepal and Tanzania,
ability was high. imported blended food had also been distrib-
If refugees perceived the food item to be uted interchangeably with the local product. In
included in the ration at the expense of the both cases, there was a preference among the
reduction or removal of another commodity, refugees for the imported product, but the local
this reduced its acceptability. Familiar foods product had not been rejected. There was no
and cooking methods were largely maintained, facility in-country to analyse samples for their
but not to the exclusion of innovations which micronutrient content; thus a key aspect of
were understood to be beneficial, such as quality-control remained problematic.
parboiled rice, blended food, yellow maize The study did not reveal major problems
meal, and improved stoves. with either use or acceptability of blended
foods, but did highlight some technical and
operational issues of quality-control and timely
Fortification of cereals with supply of the locally produced products. As
micronutrients expected, the strategy of cereal fortification was
shown to involve major issues of technical and
The second part of the study, related to cereal operational feasibility in the two African sites.
fortification, found that regional-level fortifi- However, the study revealed aspects of cereal
cation would require adequate milling capacity use and acceptability which would also need to
relatively close to the refugee population; a be considered for successful implementation.

ACC/SCN Administrative Committee on Co-or- MR mortality rate
dination/Subcommittee on Nutrition MT metric tonne
ADFL Alliance of Democratic Forces for the MUAC mid-upper arm circumference
Liberation of Congo-Zaire
NFE non-formal education
ARRA Administration for Refugee and
NGO non-government organisation
Returnee Affairs (Ethiopia)
NMC National Milling Corporation
BFP banket feeding programme (for <5years)
NR Nepali rupees (currency)
BHU Basic Health Unit
NRCS Nepal Red Cross Society
CDC Centers for Disease Control, Atlanta USA
PEM protein-energy malnutrition
CMR crude mortality rate
PPD per person per day
CSB corn-soya blend
PRO Protracted Refugee Operation
DRC Democratic Republic of Congo
RARP Refugee Areas Rehabilitation
DSM dried skimmed milk
Programme (Nepal)
EB Ethiopian Birr (currency)
RCU Refugee Coordination Unit (Nepal)
EDP extended delivery point
RDA recommended daily amount
EMOP Emergency Operation
RGB Royal Government of Bhutan
EU European Union
RNIS Reports on the Nutrition Situation of
FAM Food Assessment Mission Refugees and Displaced Populations
FDP final delivery point RRC Relief and Rehabilitation Committee
FDRE Federal Democratic Republic of Ethiopia RVVF Refugee Women's Forum (Nepal)
IIA Hartisheikh A camp (Ethiopia) SCF-UK Save the Children Fund, UK
HFEA Hpusehold Food Economy Analysis SFP supplementary feeding programme
HIS Health Information System SNM Somali National Movement
I IMG His Majesty's Government of Nepal TFNC Tanzania Food and Nutrition Centre,
IIPB high-protein biscuit Dar es Salaam
IDA iron-deficiency anaemia TFP therapeutic feeding programme
IDD iodine-deficiency disorder TRCS Tanzania Red Cross Society
IFRC International Federation of Red Cross Tsh Tanzanian shillings (currency)
and Red Crescent Societies UNHCR United Nations High Commissioner
INGO international non-government for Refugees
organisation UNICEF United Nations Children's Fund
KB Kebribeyah camp (Ethiopia) VAD Vitamin A deficiency
MCH maternal and child health WFH weight-for-height
MDD micronutrient-deficiency disorder WFL weigh t-for-length
MHA Ministry of Home Affairs WFP World Food Programme
MOH Ministry of Health WHO World Health Organisation
MOU Memorandum of Understanding WSB wheat-sova blend

1 Main report

Introduction There is a strong link between malnutrition
(acute wasting, oedema, and micronutrient deficits)
The broad aim of this study was to contribute to and infection; severe malnutrition increases the
current discourse on micronutrient deficiencies incidence, duration, and severity of infection,
in the diet of refugees by providing some field- while infections frequently precipitate malnutrition.
based evidence regarding use and acceptability The significance of micronutrient deficien-
of cereal-based foods in relation to fortification. cies has increased in recent years, with the
The ways in which refugees prepare cereal and realisation by the aid community that 'The
blended food1 provided in the ration, and how frequent inadequacy of vitamins and other
they might perceive it, are often cited specula- micro-nutrients in refugees' rations is a major
tively as potential limitations in the implemen- cause of disease and death among them' (Keen
tation of fortification strategies. But why are micro- 1992:73).Two international forums in the early
nutrient deficiencies so important, and where 1990s2 discussed the issue and specific implica-
do use and acceptability fit in with the economic tions for policy and practice. A consensus
and technical elements of prevention strategies? developed that micronutrient deficiencies
A diet deficient in micronutrients results in constitute a major public-health issue and that,
actual micronutrient-deficiency disorders (MDDs) where refugees are dependent on international
such as anaemia and scurvy, growth failure, and aid agencies for basic food needs, the provision
also increased severity of infectious diseases. of essential micronutrients is intrinsic to
Certain MDDs, namely vitamin A deficiency maintenance of their health and well-being.
(VAD), iron-deficiency anaemia (IDA), and The failures in provision in recent years show
iodine-deficiency disorder (IDD), are endemic that such protection has not been given. The
in many poor countries. Refugees may well summary of malnutrition and MDD in the three
arrive in the host country with high prevalence field-study sites (Table 2) illustrates this.3
rates of one or all three, depending on the degree Various solutions have been proposed to
to which they are endemic and the privations address micronutrient deficiencies in the ref-
suffered during earlier displacements and the ugee ration. The most obvious solution is to
period of flight. However, in recent years, some promote access to micronutrient-rich foods by
refugee populations in Africa and Asia have other means, such as encouraging refugees to
suffered outbreaks of MDDs as a direct conse- cultivate their own food, but this is often
quence of being in refugee camps; these politically or practically not feasible — and if it is
outbreaks have included MDDs which previously feasible, it is a solution for the medium to long
had been virtually eradicated, such as pellagra term. Provision of fresch foods is the obvious
and beriberi (ACC/SCN 1995). These conditions, alternative, but practical constraints of procure-
if not recognised and treated, can mean serious ment, transport, storage, and handling limit
illness, disability, and in some cases death. this option. Fortification of foods, particularly
Micronutrients have been categorised as the cereal which is usually the staple food and
either Type I or Type II nutrients, according to therefore consumed by everybody, is perhaps
the body's response to deficiency (Golden the option with the greatest potential. However,
1991). Deficiency of Type I nutrients results in to date the provision of fortified foods in the
specific clinical disorders, while deficiency of general food ration has been limited to blended
Type II nutrients results in growth failure. An foods, oil, and salt.
adequate and balanced intake of both types is The most recent Memorandum of Under-
thus required to maintain nutritional status, and standing (MOU) between the World Food
to enable nutritional rehabilitation. Absence of Programme (WFP) and the United Nations
clinical disorders, however, does not mean High Commissioner for Refugees (UNHCR)
absence of deficiencies. includes the routine provision of a blended food

Acceptability and use of cereal-based foods in refugee camps

in the ration: 'Where beneficiaries are totally Structure of the report
dependent on food aid, WFP will ensure the The following section presents the methodology
provision of blended foods or other fortified for the three case-studies. This is followed by a
commodities in order to prevent or correct review of important contextual details, including
micronutrient deficiencies' (WFP/UNHCR 1997). the recorded outbreaks of micronutrient defi-
The amount of blended food included in the ciency, and external factors that influenced
ration is planned to be between 30 and 60 whether refugees actually received what was
gm/ppd, depending on factors including the planned. The next section considers the refugees'
overall composition of the ration (Bhatia 1997). preferences for different foods in the ration, and
The same MOU has a commitment also to the preferences between ration-food items and
providing milled grain rather than whole grain, non-ration food items of the same type. We then
especially in the early stages of an emergency; if examine the actual use of ration foods by the
whole grain is provided, local milling capacity refugee women, and go on to identify and explore
must be available, and the ration should include the wide range of determinants of acceptability.
compensation for milling costs where these are The next section explores the feasibility of
borne by the beneficiaries (WFP/UNHCR 1997). cereal fortification at the regional, community,
and household levels. The final section presents
Why consider use and acceptability?4 the main conclusions from the study.
The rationale for providing blended food as a The details of the methodology and the three
strategy to combat micronutrient deficiencies case-studies are presented in full following the
has been in doubt, partly because of a lack of main report. Throughout the text, the studies
understanding and documented knowledge are referred to according to the host country
about refugees' actual use of blended foods and visited, i.e. Nepal, Ethiopia, and Tanzania.
their acceptability to the recipients.
For example, anecdotal reports have suggest-
ed that blended food is used and perceived by
refugees as a food for children, and as such that it
is ineffective as a vehicle for boosting micro- The study was designed to generate infor-
nutrient intake of all groups in the population. mation about how particular groups of refugees
However, there is little documentation of the used the food available to them, the accept-
food habits of refugees, and less is known about ability to them of predominantly cereal-based
their actual use of blended food, in terms of who food in the refugee ration, and the possible
is eating it, how it is prepared, whether it has a opportunities for cereal fortification.
resale value, and how it is perceived by the
The refugee populations chosen for study
were located in geographically diverse regions
In contrast, the validity of a cereal-fortification (Nepal — South Asia; Ethiopia — East Africa;
strategy seems to depend more on questions of and Tanzania — Central Africa), but they
technical and operational feasibility than use and shared a common dependence on international
acceptability of cereals, but this too is considered relief which meant that they were, or had been
to be important, for example in deciding at which recently, nutritionally vulnerable.
point cereals could and should be fortified: either The choice of suitable case-study sites was
before, during, or after distribution (Beaton limited, partly because of the practical constraints
1995). on undertaking research of this nature at the
In order to inform the debate, this field study epicentre of an emergency, not least of which
aimed to investigate the aspect of use and was the unpredictability of acute displacements,
acceptability of both blended food and cereal, and the personal security of the researcher and
and opportunities for cereal fortification in three interpreter in potentially unstable situations.
sites, studied in 1997, as follows: This, combined with the need for sites to meet
1. Bhutanese refugees in Nepal: all camps specific criteria (see Appendix l(a)), restricted
(1992 to date). the availability of suitable sites. The third site
2. Somali refugees in Ethiopia: Kebribeyah was not confirmed until half way through the
and Hartisheikh A (1988 to date). research, which illustrates the difficulties of
3. Burundi refugees in Tanzania: Muyovosi finding appropriate locations for field research.
camp (1996 to date). The total period in-country for each study
was approximately seven weeks during 1997.

Main report

Nepal was visited in January/February, Ethiopia interests vested in the system. To try to avoid
in May/June, and Tanzania in August/ September. raising expectations which could not be met by
The choice of study sites and other aspects of the field study, research methods involving
methodology are described in detail in the case- large or formal groups were mostly avoided.
study appendices. The same researcher was Inextricably linked with this fact was the
involved in all three studies, assisted only by a extremely sensitive issue of numerical estimations
locally recruited female interpreter who had no of populations, and entitlements based on these.
previous connections with either the refugee This affected the fieldwork and analysis. Inform-
population or the relief programme. ation concerning amounts of food obtained, sold,
A research protocol was developed in advance, and bought was influenced by the refugees'
describing in detail the methodology for the perceptions of the researcher as being in a
study. The same methods were used for each of position to influence what food might be
the three studies, which included the collection supplied, and in what quantities. The researcher's
of a combination of primary ethnographic data attempts to explain the purpose of the study and
and secondary data from each study location the role of the commissioning agency probably
(see Appendix l(a)). The primary data were meant very little. While the attitude of individual
obtained through key-informant interviews, obser- refugees was mainly welcoming, in Ethiopia some
vation sessions, and household interviews. A women were overheard to say within earshot of
range of key informants was chosen, including UN the researcher: 'Don't bring her to my house'.
and NGO personnel in the capital and at regional This was partly due to 'survey fatigue' on the part
and camp levels; refugee women's representatives, of the refugees, but also reflected the general
and other refugees with particular knowledge; atmosphere in this study site, an issue which is
factory managers, mill owners at camp and local explored further in the case-study and
levels, and a milling consultant (Tanzania). mentioned later in this report.
The secondary sources of information included Refugees were often defensive when dis-
a wide range of mainly unpublished UN and cussing amounts sold or exchanged, because they
NGO material. This were particularly useful in thought the agencies did not approve of such
providing contextual information which could activities. This reticence meant that intra-
then be discussed with key informants. household distribution was discussed in interview
The findings from the three case-studies but not observed formally.
were exploratory, and cannot be representative
Security: One criterion governing the choice of
of all refugees in the study sites. The three case-
study sites was that the situation should be
studies examine qualitative relationships
reasonably stable and secure. This meant there
between food practices, refugees' perceptions,
was some delay in beginning the first field study,
and existing conditions related to rations and
which was planned to be in the Great Lakes
other sources of food. The study was not area; but there were major upheavals in that
designed to obtain quantitative data on the use region at that time, so that it was decided to post-
of food, although some secondary quantitative pone the visit there and quickly re-arrange the
data relevant to local conditions were collected. schedule so that thefirstfieldstudy was in Nepal.
Methodological constraints In the two sites in Africa, the camps were one
hour's drive from the researcher's base, and
There were two major constraints that affected agency staff were encouraged to leave the camp
the process of data collection and in some cases by mid-afternoon (Ethiopia) and before dark
may have had a bearing on the nature of (Tanzania). Thus most field work took place in
collected data. These constraints were food- the morning hours. In Tanzania the researcher
related sensitivities in refugee situations, and the was told to keep a radio handset at all times
problem of security. when visiting in the sections, because of recent
Food-related sensitivities: Most of the sensi- outbreaks of tension between refugees.
tivities of this study stemmed from the fact that
in refugee contexts food is the largest single
resource available to refugees. Food assistance Cereal fortification and blended
represents the biggest transfer of resources food
from the international aid system to refugees
overall. Thus in the case of both refugees and 'Food fortification may be broadly defined as the
agency staff, responses were influenced by process whereby one or more nutrients are added to

Acceptability and use of cereal-based foods in refugee camps

foods to maintain, improve or enhance the quality of a depending on whether the food is for supple-
diet consumed by a group, community or population mentary feeding or for inclusion in the general
(Henry and Seaman 1992). ration. Briend (1989) recommends the develop-
'Cereal fortification' in this study refers to the ment of two different formulas for these two
fortification of the main cereal staple in the different purposes.
refugee ration with essential micronutriqnts by
the mixing in of a vitamin/mineral pre-mix at
the time of milling grain to meal.6 This could in Food and nutrition in the study
theory take place before or after distribution to sites
the refugees, but for practical reasons of cost
and quality control it has been considered The three field studies illustrate the complexity
preferable to fortify in bulk, i.e. before of refugee situations and their diverse range of
distribution to refugees, and to be possible only characteristics (see Table 1). The unique
if the necessary skills and machinery are made character of the local context, including the
available in the host country, or if the food-aid humanitarian response, exerts a considerable
donors mill and fortify food prior to shipping, influence over the acceptability and use of food
which has implications for shelf-life (Henry and and is therefore worthy of review. This section
Seaman 1992). To date, cereal fortification has also provides an overview of the recorded MDD
been implemented only once on a large scale in outbreaks, and the ration composition in each
Africa: for refugees in Malawi, maize meal was of the three sites. This is followed by a review of
fortified with B vitamins following an outbreak the more important external factors which
of pellagra in 1990. influenced whether refugees actually received
Pre-cooked, fortified blended foods are a what was previously intended. The section
mixture of a cereal and oil seed (usually soya), closes with a review of the systems of food
and sometimes pulses (such as chickpea), which distribution and monitoring, which considers
have been milled and cooked by either roasting issues of gender and access to food, and also
or extrusion, and a vitamin and mineral mixture refugees' perceptions of the system.
added. Those available in refugee situations are In all three sites, the planned ration was more
usually either produced in the USA (corn-soya than 1,900 Kcalories ppd, which indicates that
blend (CSB) and wheat-soya blend (WSB), for the agencies recognised that the refugees were
example) or locally manufactured. In the follow- dependent on food assistance as their main
ing text, the term does not include foods fortified source of food. All three situations could be
with a rich protein source, such as soy-fortified categorised as 'nutritional emergencies' according
cereals (for example soya-fortified corn meal). to the response by agencies.
Blended foods have been supplied for many There is no doubt that these refugees were
years to refugee populations for use in supple- nutritionally vulnerable, but whether the situations
mentary feeding programmes. In recent years corresponded to a 'nutritional emergency' dep-
they have been included in the planned ration ends how that is defined: whether by response
for those dependent on food aid, to prevent or or outcome. If 'nutritional emergency' were to
correct micronutrient deficiencies. One of the be defined only according to rates of acute
main issues arising from this strategy has been malnutrition,7 i.e. by outcome, then at the time
that of cost. The blended foods are mostly of the study this categorisation of Nepal and
produced in the USA (CSB and WSB), but Tanzania would be questionable.8
locally produced blended foods have also been Nepal and Ethiopia were situations in the
procured by donors and production facilitated operational 'care and maintenance phase' as
by WFP. It is extremely difficult to compare defined by UNHCR, while Tanzania was still in
costs of blended food and alternative strategies the 'emergency phase'. Similarly, in terms of
such as cereal fortification, which has rarely WFP's classification of phases, Nepal and
been implemented to date. Blended food ex- Ethiopia constituted a PRO (protracted refugee
USA is relatively expensive (Appendix l(c)); operation), i.e. they continued to be dependent
but, as a vehicle for micronutrients in the ration, on food assistance beyond the initial 12-month
the amount required is proportionately much period; Tanzania constituted an EMOP (emer-
less than that of the staple cereal. Another issue gency operation), with provision of food for the
related to blended food is its micronutrient first 12 months. These classifications also relate
composition, and whether this should vary to funding issues and procedures in both WFP

Main report

Table 1: Study proposal criteria and field-study sites

Study Criteria Nepal Ethiopia Tanzania
General ration (planned) set 2,285 Kcals 1,840 Kcals 1,950-2,100,depending
at minimum 1,900 Kcals ppd whether grain or meal

Blended food included in Yes —since 1994 1988-1992—erratic Yes
the general ration supply
1992-1997 —no
1997 — erratic supply

Limited access to local Good access 1988 — very poor Good access
markets access, then markets
in camp vicinity + +

Few income-earning Casual labour, restricted Variable depending on Casual labour,
opportunities movement > daily clan links, number of restricted movement
ration cards.
Major and petty trade

Limited access to land for Extremely limited Limited land for Very limited
cultivation cultivation and grazing

Large refugee population Yes (Beldangi) Yes Yes
relative to local population

Table 2: Summary of malnutrition and MMD in the three case-study sites

Nepal Ethiopia9 Tanzania
Condition pre-flight No data, assumed No data, assumed High rales of acute
reasonable reasonable malnutrition in areas
affected by violence

Condition on arrival Some deterioration Reported good in 1988 High rates of acute
during flight malnutrition

The first 6 months High prevalence of acute High prevalence of acute High rates of acute
malnutrition malnutrition malnutrition
IDA main MDD concern Scurvy prevalence 1-2% IDA ++ (clinic data, not
within 6 months survey, cause attributed
to malaria)

6-12 months Declining rates of acute Acute malnutrition Rapidly declining rates
malnutrition High prevalence of scurvy of acute malnutrition
IDA ++ (survey results
yet to be confirmed)

12-24 months Low rates of acute Acute malnutritionrateslow n/a
malnutrition Scurvy rates declining,
Beriberi outbreak 18 c.30 cases/month in
months after main influx TFP/SFP

After 2 years Low rates of acute Moderate-high rates of n/a
malnutrition acute malnutrition
Low levels of vitamin B Individual cases but no
(and C) deficiency significant outbreak of
reported intermittently to scurvy or other MDD to
date, cause unknown. date

Acceptability and use of cereal-based foods in refugee camps

and UNHCR. Hence if'nutritional emergency' health staff. This may partly account for on-
is defined by phase rather than outcome or going reports of low incidence levels to date.
responsc, then of the three study sites only In Nepal, the beriberi outbreak occurred
Tanzania would be included . when acute malnutrition rates were low. The
Outbreaks of MDD have been a feature of general ration was supplied regularly, but at the
several refugee situations in the last ten years, time of the outbreak it included polished rice,
and have occurred in both the 'emergency' and and no blended food.
'care and maintenance' phases, at times of both In Ediiopia, the nutritional status of the people
high and low levels of acute malnutrition. A arriving as refugees in 1988 was good. Scurvy
review of outbreaks over the years illustrates developed within six months of the main influx.
this (Table 2), as does the history of mal- In thesefirstmonths, there was virtually no alterna-
nutrition and micronutrient deficiencies in the tive food supply to that provided in the ration,
three study sites summarised in the same table. which was erratic and incomplete. Blended food
Furthermore, refugee situations are not always was supplied, but often as a substitute for grain.
characterised by a linear continuum from non-
refugee to refugee status, with clear distinctive Micronutrient content of the planned
boundaries to identify each phase along the ration
continuum. The epidemiology of nutritional The 'planned ration' in Nepal and Ethiopia was
disorders reflects this complexity. The refugee reviewed and changed accordingly over the
case-studies included in this study demonstrate years. In all cases the planned ration was provided
cyclical patterns (Ethiopia) and multi-stage as a nutritional intervention, i.e. as a source of
patterns (Ethiopia/Tanzania) of refugee move- essential nutrients, not as an income transfer or
ments. In Ethiopia, many of the camp residents economic support. Details of the ration compo-
had previously been refugees in camps in sition are given in Appendix l(d), together with
Somalia, and among the other camp residents the percentage of the WHO-recommended
were people who had arrived from Somali towns nutritional requirements met by each ration.
in 1988, from rural areas of Somalia in 1991, and The planned ration scales all show consid-
from many areas in 1994. Some were Somali erable shortfalls in the micronutrient content,
citizens; others were returnee Ethiopians; there which were not fully compensated for by the
were also locally displaced people. Some had addition of blended foods.
spontaneously settled or repatriated, some had
done so and then returned. In Tanzania, many of Did the refugees get the planned ration?
the refugees had already spent several months As most field workers know, there is a world of
displaced inside Burundi, while others had been difference between the ration that is carefully
refugees for up to three years in Zaire. planned and calculated and the food that is later
When interpreting Table 2, it is important to received by beneficiaries. In all field sites, there
consider that reporting protocols were not had been problems in ensuring that refugees
standardised, diagnostic competency was not received the planned ration in full and
uniform, and for vitamin B and vitamin C regularly. There were, to a greater or lesser
deficiencies case-definitions were not well estab- extent, considerable shortfalls from the amounts
lished. While under-reporting was thus often a originally planned and agreed. It is WFP policy
consequence, there may also have been not to provide retroactive issues of food to
situations of over-reporting through mis- refugees when there have been gaps in supply,
diagnosis or flawed surveillance techniques. so these shortfalls remained uncorrected.
The absence of MDD does not indicate The reasons for gaps in supply and shortfalls
unequivocally adequate levels of micronutrient in rations received are manifold and often
intake. MDDs other than VAD and IDA were context-specific. However, the more important
not recorded on the health centres' reporting causes can be grouped as follows:
forms (except in Nepal), and the capacity for • Restricted access to the affected population
accurate diagnosis was variable. Also variable for reasons of remote locations, insufficient
was the level of awareness of MDDs, depending infrastructure (roads, transport networks,
largely on people's experience of them; so that, etc.), seasonal closures, and possible
for example, in Nepal after 1994 there was a insecurity.
high sensitivity to symptoms associated with • Lack of resources and variable donor
vitamin B deficiency among both refugee and commitment.

Main report

• Disagreement over accuracy of beneficiary and implementing agencies and donors.
numbers linked with registration. In Ethiopia, there was doubt about the actual
• Erratic distribution system. size of the refugee population in the camps
• Erratic monitoring of distribution and because of a long history of poor registration
complaints. and the proliferation of ration cards, many of
them forged. Some refugees had several cards,
These are not listed in order of magnitude or
while others had only one. The most recent
significance, but rather trace the food from
verification of numbers had been in 1994, just
source to beneficiary. They are not all mutually
prior to another influx, since when there had
exclusive. An awareness of these problems is
been no further verification. To complicate matters
necessary to understand the external factors more, the camp residents included refugees,
which influence both food availability within the returnees, and some locally displaced groups
refugee households and also food use. who did not share the same entitlements (accord-
Access: There were logistical impediments to ing to UNHCR) but who nevertheless continued
supplying the planned ration to the refugee to co-habit in the camps. Inflated numbers and
sites. Access was most difficult in Ethiopia, the lack of a verification in 1996 had further
where, in the early days particularly, there had undermined the credibility of appeals for aid,
been serious problems caused by the remote- despite other sources of information indicating
ness of the site. Roads had had to be repaired, pockets of need and apparently persistent high
and maintenance is still on-going. In this area, levels of acute malnutrition in some camps.
too, there had been several periods of insecurity Thus poor registration had a direct impact on
and banditry over the years.10 donor commitment and response.
In Tanzania, there had been and continued At the time of the field studies, in Nepal there
to be problems of damage to local rural roads, was no great discrepancy between registered
which were not built to take the 40MT-capacity and actual numbers, and verification was being
lorries required to serve the camps. The wet undertaken at the time of the study. The
season brings major delays, so there is a need to sensitivity here was the connection between
stockpile 2-3 months' advance supply at the registration and refugee status in relation to
extended delivery points (EDPs); this was being rights of citizenship on return to Bhutan. In
undertaken at the time of the study. Tanzania, a registration had been undertaken
In Nepal there were no logistical problems, two months prior to the study, and demographic
except for the building and maintenance of data entered into a computerised database to
local feeder roads. The food pipeline was short enable on-going verification. There continued
and transport abundant. to be small numbers of'new arrivals', many of
whom after assessment were said to be 'recyclers'
Resources and donor commitment' If the
and so not eligible for another ration card.
resources are not available to buy the necessary
food aid, the programme cannot proceed as Distribution system: Details of the distribution
planned. In Ethiopia unmet appeals for resources system are important, not only in ascertaining
for the programme meant that WFP was often bor- how much of dieir entitlement refugees received,
rowing cereal locally, in a 'hand to mouth' fashion. but also in assessing to what extent they could
These local loans required proof that a shipping trust in the timeliness and content of the distrib-
order had been made, and they had to be repaid ution, which was critical for household-level
in full. It was impossible to build up a buffer stock, planning and management of resources. Incomp-
and this in turn influenced the feasibility of lete and unpredictable distributions constituted
regional-level milling and fortification of cereal. a major grievance among refugees in Ethiopia,
In contrast, food supplies in Nepal were where there were gaps of between four and six
assured, and other sectors, notably education, were weeks in the monthly distribution — in contrast
well supported in the camps. In Tanzania a to Nepal and Tanzania, where distributions
commitment to the population and the wider were timely and regular (every two weeks).
regional situation was demonstrated by the
All three study sites had a system of
building up of buffer stocks, and the up-grading
of roads, ports, and logistics bases. distribution to groups of heads of families
(rather than through group leadership, or to
Registration of beneficiaries: Registration and individual heads of family). However, all three
on-going verification of numbers had been a distributions looked quite different. All systems
highly contentious issue in all sites, for refugees purported to involve the refugees. The criteria

Acceptability and use of cereal-based foods in refugee camps

for grouping refugees in Nepal were based on refugee households. These studies suggested
'sub-sectors' (each sub-sector comprising the that the handing over of the ration to women at
families within it, all of different sizes), while in the distribution site did not influence how much
Ethiopia and Tanzania the groups were formed they controlled it beyond that point.
on the basis of household size (each group
comprising families all of equal size). Monitoring of distribution and complaints:
There was formal on-site food-basket monitor-
The main contrasts were evident in the level
ing only in Tanzania, and this was not being
of organisation at the distribution site, as well as
conducted competently at the time of the study.
the level of transparency and availability of
There was no post-distribution monitoring at
information about where and how power over household level, and regular market-price
the food resided. In Nepal and Tanzania, surveillance was being done only in Tanzania.
WFP/UNHCR had considerable power over the
Women in all three studies were notably
resource distribution, and there were formal,
reluctant to make any criticism of the distri-
regular meetings with implementing agencies
bution system, even in the privacy of their
and refugee representatives. A number of
homes in one-to-one interviews. This applied
parties were involved. In contrast, in Ethiopia
even in Ethiopia, where the system was most
such meetings were not open to scrutiny, and
open to abuse and misappropriation. Where
there was only one implementing partner refugee participation was encouraged and visible,
involved for all sectors (excluding water-tanker- complaints were not necessarily more forth-
ing), with no monitoring of the distribution by coming than in the closed system. In Nepal
any party. This allowed scope for misappropri- recipients might be reluctant to complain
ation and meant that improvements in the because it meant openly mistrusting a neigh-
distribution system were unlikely to be initiated. bour, i.e. someone from the same sub-sector. In
The Food Assessment Mission (FAM) in 1996 Tanzania, similarly, it was said that it was not
stated that 'this system affords no assurance that appropriate to complain about the group leader,
(1) full rations reach the intended beneficiaries i.e. the person who can be and is of assistance.
or (2) group leaders represent the families that
they have registered' (WFP/UNHCR 1996a). While refugees might be reluctant to
complain about the distribution system, they
In Ethiopia and Tanzania, men were pre- invariably complained in the interviews that the
dominantly in evidence as the initial recipients food brought for them was not enough, and
of the food, i.e. as group leaders. Women were almost always also that it was monotonous, and
comparatively most evident at this level in that there were foods which they would prefer.
Nepal. In all sites, there was no apparent Before looking in detail at use and acceptability
restriction on any person within the household of the food given through the ration, it is useful
collecting the ration from point of distribution, to consider what it was that the refugees would
so that men and women and children, depend- have preferred, and why.
ing on who was available at the time, would
collect and bring the food back to the house
(observed and confirmed in interviews). The 'pragmatics of preference'
However, although women could receive the
ration and bring it to the household, this did not During interviews, a simplistic question about
mean that they continued to control the ration food preference was asked which sometimes led
thereafter. In the house, women were the ones to confusion, especially if couched in a hypo-
responsible for allocating amounts and deciding thetical way. In one interview in Ethiopia, a
what and how much to prepare, but they did not woman observer interrupted and suggested a
have the same level of responsibility over useful distinction between what the woman
deciding what portion of the ration went out of refugee might prefer now in her present situation,
the house for sale or exchange. This division of and what she would prefer if not so constrained.
responsibility was most evident in Tanzania. In Consequently the concept of preference has
Ethiopia, women appeared to have more control, been broken down as follows:
partly because of the common prolonged
absence of men from the camp, and partly Preferences within the range of ration
because of the economic control which Somali food items only
women have traditionally exercised as traders In Tanzania there was an intriguing hierarchy
in their own right, and latterly as heads of of preferences regarding cereal. White maize

Main report

grain and yellow maize meal were highly special cases (illness, religious ceremonies).
acceptable, some interviewees favouring one However, there was a preference for halo dual
over the other, but many favouring a combi- (not in the ration); but this was much more
nation. However, yellow maize grain was less expensive than the chana, so the latter was sold
favoured, and white maize meal was most to buy more vegetables, essential for palatability.
unpopular. The factors at work here were Cassava flour (not in the ration) was preferred
mainly resale value, milling costs, and quality of for the making ofugali in Tanzania, but it was
product. In Ethiopia the hierarchy of preference relatively expensive, so mostly the ration maize
of cereal was wheat, sorghum, and lastly maize; meal was used, often mixed with a smaller
market value and versatility were the determining proportion of cassava flour. Thus a choice based
factors, not ease of preparation or cost of on familiarity was often constrained by purchas-
milling. ing power. One notable exception to this was
In Ethiopia there was said to be a preference the case of oil exchange in Tanzania, where
for wheat over sorghum, but not all inter- locally produced palm oil was obtained for twice
viewees confirmed this. Here the preference for the amount of ration oil.
wheat was influenced by its versatility and also
by its market value, and was probably felt more Preferences within the range of items
strongly by those who sold in bulk. Some key formerly available
informants from agencies had reported that the This range of preferences indicates the tastes
sorghum distributed in the previous year had established before people became refugees. In
been deeply unpopular, but this was not all cases, predictably, most people had formerly
overwhelmingly supported in interviews, and enjoyed access to a much wider range of food
some women, from pastoralist backgrounds, commodities, at least before the disruptions
expressed a preference for sorghum over which led to their displacement, and in Somalia
wheat. Opinions differed among informants, and Burundi their common staples had not
with those who wanted to see wheat included in been those which were now part of the general
the ration because of its resale value possibly ration. In Bhutan, the basic daily fare had been
exaggerating the poor quality of the sorghum. very similar to items being distributed in the
While most refugees in 1994 in Nepal pre- ration, but in addition people had produced
ferred polished rice to parboiled rice, by 1997 a and eaten a much wider range of vegetables,
significant number were saying in interview that and also dairy products. In Somalia, urban
not only had they accepted parboiled rice, but dwellers had purchased most of their food, and
that they now actually preferred it. Explan- factory-refined flour, bread, rice, and pasta had
ations for this included digestibility and the fact been widely available, while in pastoralist house-
that now it did not smell as it had done at the holds meat and milk were important staples,
beginning. (This has implications for quality of and for both urban and pastoralist households
the processing.") In addition, many associated tea with milk and sugar was a significant part of
it with improved health status generally. the daily fare. In Burundi, instead of tea,
Where more than one type of blended food indimano, the traditional weak beer, had been
had been distributed (Nepal, Tanzania), the imbibed most days. In terms of staples, the
imported USA product (WSB and CSB strong preference was for plantains, beans, and
respectively) was preferred. It was noted that root crops such as cassava and sweet potatoes as
packaging of the local product was inferior in daily staples, rather than grain. They had been
Tanzania, but this was not mentioned by inter- largely subsistence farmers, so had grown most
viewees. Locally produced blended foods were of what they ate. The most prolific crops had
said not to be ground very well, and in Nepal to been cassava and plantain, potatoes and beans,
smell of maize. Here the perceived inferior not maize or any other cereal.
quality of the content and texture was the main Thus established tastes and preferences ensured
basis for preferring the imported product. a wide source and variety of micronutrients.
Now they persisted, but were modified to some
Preferences between ration food and non- extent by newly acquired tastes and a prag-
ration food items of the same type matism born of the constraints of affordability
In Nepal, as discussed above, there was no and availability.
strong preference for polished rice over How were these competing elements
parboiled rice (not in the ration), except for managed by refugees?

Acceptability and use of cereal-based foods in refugee camps

The use of food and lentils were on sale after they had been
micronutrients distributed as part of the repatriation packages.
These were being bought by women who were
Food was managed by refugees at a number of not repatriating and were using the lentils to fill
levels and in a complex range of settings, which the samosas they were making for sale.
were inter-related. There was the 'marketplace' In Ethiopia, again in contrast to the other
outside the camp boundary, at a district or sites, there was obviously quite widespread bulk
regional level; there was also the 'marketplace' selling. Holders of multiple ration cards were
within the camp, in addition to inter-household abusing the system, but their activities were not
arrangements and activities within the house- investigated for this study because the business
hold. Thus the 'marketplace' included social interests involved were not those of refugees
and economic relationships and was not largely dependent on the general ration. Apart
confined to the physical space of the market. from this large-scale business, there was bulk
There were restrictions in place in Nepal and selling by refugees, a fact openly admitted in
Tanzania on the movement in and out of camps interviews. Men and women collecting the
ration were hiring wheelbarrows and taking
by refugees and non-refugees, for a combination
their full sacks directly to the market nearby.
of political, economic, and ecological reasons.
The money obtained was often used to pay off
They were not 'closed camps' in terms of
debts incurred over the previous weeks from
physical barriers, nor were the sanctions against
the purchase of sorghum, milk, sugar, tea, and
transgressors extensively or consistendy applied. (for some women) costs incurred in the making
Nevertheless, the limits were known by refugees of samosas for sale. In this case, a particular
and affected their mobility, so that most of them nutritional advantage was the acquisition of
were living in the camps, leaving only for daily fresh milk, which was usually added to sugared
casual work or trading. In contrast, in Ethiopia, tea, or drunk by itself, and was also added to
where there were no restrictions, there was some of the dishes made from the grain staple.
considerable traffic between camps and within It was also a complementary food for infants
the region (including across the international and young children.
border) and a marked discrepancy between
actual residence and numbers registered. Another form of bulk trading was the sale of
Relationships between the local communities 'pooled food', where a group of neighbours
and refugee populations in all sites were mostly contributed small amounts to a central pool
peaceful and mutually beneficial. over time and took turns to sell the accumulated
pooled amount and keep the proceeds. In
Sale and purchase Nepal this informal system was said to exist,
although women in interviews did not
In all sites, a proportion of ration food was sold
volunteer information, presumably because
by refugees. For those who were heavily
they feared it would be deemed a misuse of the
dependent on the ration, the money obtained
ration. Similarly in Ethiopia a co-operative
was used to purchase other food items. The
system among women, known traditionally as
scale and timing of food sales varied, but most of
hakbet, was said by a key informant to exist in the
the food seemed to be sold on or soon after
camps using ration food, but women in
distribution day to agents or directly to traders.
interviews were unforthcoming. In Tanzania,
However, in Nepal and Tanzania there was
there were suggestions of similar activity, for
some mention of traders visiting the sections at
example, lentils being collected and accumu-
other times to collect from those known to be
lated for sale and a group of neighbours
likely to want to sell some amount.
involved in 'co-operative' work together, but in
In Nepal and Tanzania, WFP and UNHCR all cases refugees were reticent in giving details.
had judged the amounts seen in the market to
be within acceptable limits and not to be Sale of cooked ration food: Ration food was also
adversely affecting the local host economy. In sold as cooked food, again to obtain cash for
Ethiopia the picture was more complex: the other items. In Ethiopia this was in the form of
camps were integral to a much wider regional samosas, using wheat flour and lentils, the latter
economy within the host country, and part of an being part not of the general ration but of the
economic continuum which included the country repatriation package. In Tanzania the product
of origin. There was also (for some) the prospect was mostly alcohol, made from a combination of
of voluntary repatriation. Large quantities of sorghum, purchased from the sale of maize

Main report

grain, and maize grain from the ration. This common to all sites, even in Nepal, where
alcohol was relatively cheap and widely used by vegetables were included in the general ration.
men and women, and, in the mildly alcoholic or In this case, the main reasons given were to
juice' phase, consumed also by children. This break the monotony of eating lentils twice daily.
activity could be considered as protective of Vegetables were considered essential for any
micronutrient content, the low-alcohol beer being meal with rice. Rice could be eaten without
'more a food than a beverage' (FAO 1995). lentils, but not without vegetables. In the other
There was no apparent cooked product on sites, vegetables were important in improving
sale in Nepal. This may partly be due to the palatability of the staple. The popularity of
traditional observances related to raw and vegetables was underlined by the cultivation,
cooked food and the hearth. Cooked food was wherever possible, of green vegetables (partic-
often shared with visitors as part of ritual cere- ularly in Nepal and Tanzania, where they are
monies; however, it was usually not food from very easy to grow), and the consumption of
the ration, but food acquired often through sale bean, cassava, and pumpkin leaves. The type of
or exchange of part of the ration; it usually vegetable obtained was partly dependent on
included polished rice, rice flour, and ghee. season, but in Nepal and Tanzania there was at
These foods were less rich in micronutrients but least one popular, cheap, and easy-to-grow green
were not daily fare; rather they were essential vegetable (saag and lengalenga respectively). In
for the fulfilment of ritual obligations. Ethiopia green vegetables were not so evident,
but tomatoes and onion were used in many
Sale of blended food: In Nepal and Tanzania dishes, and wild fruits were gathered or pur-
there was a relatively very low level of trading in chased when available. In Nepal and Tanzania,
blended food. The exception was Ethiopia, root vegetables were also very popular, such as
where the blended food on sale was mostly in potatoes (Nepal), and cassava and sweet
the form of pre-mix, and there were reasons to potatoes (Tanzania). In Burundi, these root
suppose it had come from the feeding-centre vegetables had usually constituted the staple.
store, not from individual recipients. In Nepal it
was not observed in the markets in and adjacent In Nepal, some agency staff speculated that
to the camps, but non-refugee informants said the high refugee birth rate might be explained
they bought in small quantities from individual by the fact that this was one way to bring more
refugees inside the camp. In Tanzania relatively food into the house: small children being unable
small amounts (compared with grain and meal) to eat their daily rice ration and a surplus of rice
were on sale in the camp market intermittently, thus resulting. In fact, the health-surveillance
but none was found in the district town. At the data showed that the birth rate had decreased
time of the study, prices in Nepal were lower substantially over the previous two years.
than the price of rice, and lower than maize However, it is interesting that agency staff, and
meal in Tanzania. This generally indicates a indeed some of the refugees themselves, should
nutritionally positive use of a food which is harbour these perceptions, when in the other
micronutrient-dense and for which there is no two sites, women refugees, discussing who
locally available substitute. The reasons given by might have surplus food to sell off, identified
women for not selling blended food were bachelors or families consisting mostly of adults.
similar in all sites. Firstly they said that they Non-food items were also purchased; mostly
received so little of it from the distribution that it these were fuel, household wares, and clothing.
would not be worth selling, for the resale value In Nepal the refugees received a regular fuel
was low. Secondly, they said that it was very ration, and also periodically diere were distribu-
popular with all members of the household, tions of shelter material, kitchen sets, and
particularly children, and (in Nepal and clothing. This was in marked contrast to the
Tanzania) at times indispensable as a breakfast small-scale distributions of non-food items
and snack food. In all sites, women in interview (NFI) and commitments in Tanzania and the
tended to express surprise, concern, and even non-existent programme in Ethiopia. Non-food
disapproval in some cases, when the possibility, items such as clothing and household wares
was mentioned that some people, albeit a very were least in evidence in Tanzania. Here the
small minority, might be selling it. population was relatively newly arrived, and
mostly from poor rural origins, with very little
Purchase: The purchase of vegetables with opportunity to bring goods with them when
money obtained from sale of the ration was they fled from their homes, whether they had

Acceptability and use of cereal-based foods in refugee camps

come from Burundi after periods of internal Food processing within the household
displacement, or whether they had come from The main storage problems inside the houses
refugee camps in Zaire. arose from lack of suitable containers to protect
the food from rodents and, particularly with
meal and blended food, from contamination
Exchange in place of sale was a feature in Nepal, with dust. None of the field studies was
where nearby villagers visited the camp and conducted during rains, and problems of wet
bartered on an individual basis. Here the ration and damp were not raised. In Nepal, where
item most traded was the pulses {dual), in screw-top plastic jars were available in local
exchange for vegetables. This has led to markets, and people had brought lidded tins
concerns in the past, and even currently, that with them from home, the blended food was
the (albeit very low) incidence of vitamin B well protected. However, in the African sites,
deficiency still being reported is a consequence blended food was mostly kept in plastic bags
of this trade. suspended from the ceiling of the huts, or in
Exchange rather than sale of oil was a feature sacking. One reason why grain was milled a little
in Tanzania. The vegetable oil given in the at a time was because grain is so much easier to
ration was exchanged by many almost immedi- store than meal or flour, which, if spilt or
ately after distribution for local palm oil, the exposed to dust, is very difficult to recover or
exchange rate at the time being 2:1 (vegetable clean. No storage method observed was
oil:palm oil). Nevertheless this exchange detrimental to the micronutrient content of the
constituted a net increase in vitamin A content. blended food; given the circumstances, women
In all sites, women mentioned casual arrange- were taking all measures to protect it as
ments with friends or neighbours whereby small effectively as possible. In Tanzania 'leftover'
amounts of food were loaned or given, especially cooked food was kept sometimes overnight, to
to families with children. In all sites, reference be eaten for breakfast the next morning. There
was made on occasion to 'helping each other'. was no evidence in any site of cooked blended
Small-scale borrowing between households was food being stored overnight in this way. It was
also common, and would involve any of the mostly prepared in the mornings, not the
commodities, blended food being often evenings, and could be made relatively quickly
mentioned. A distinction was made in Ethiopia at short notice.
between a gift, which was unsolicited, and a There were no preparation methods observed
loan, which was solicited and granted on the which seemed to be more than usually
understanding it would be paid back. Formal detrimental to the micronutrient content of the
credit arrangements with local businesses were ration food, except multiple washing of rice
common in Ethiopia, where a number of women (Nepal), and sieving of blended food (Tanzania).
were involved in petty trade and where there When parboiled rice was introduced in 1994 in
were irregular intervals between ration distrib- Nepal, a public information campaign had
utions. There were also inter-household stressed the importance of washing the rice only
arrangements for 'pooling' food among neigh- once, in cold water, which should be saved to
bours, as described above, in all sites. add to die curry. However, subsequent enquiries
To summarise, there seemed to be three and responses during this field study indicated
broad reasons, not mutually exclusive, for the that many women continued to wash the rice
acquisition of foods in addition to, or in two or three times, and in warm water, often
exchange for, those in the ration. These were as throwing the water away if the rice had been
follows: dirty or smelling. This practice is less detri-
• for palatability and flavouring, most notably mental to the nutrient composition of parboiled
vegetables (all sites), milk (Ethiopia), and rice than to that of polished rice. Nevertheless,
palm oil (Tanzania); some health staff cited this practice as a
• for an increase in variety, to break the contributing factor in the continuing low level
monotony, such as tealeaves and sugar incidence of vitamin B deficiency.12
(Nepal, Ethiopia), fish and cassava Sieving of blended food was observed only in
(Tanzania); Tanzania, where it was done by only a small
• for special occasions, such as meat (all sites), minority of women, who began to sieve it when
ghee and rice flour (Nepal). they received a consignment of locally produced
blended food which to them seemed not to have

Main report

been ground finely enough. Some threw away than firewood (kerosene in Nepal) and different
the remainder, some kept it and later pounded types of stove (Nepal and Tanzania). In both
it and sieved again. The grain in Ethiopia and these cases, women had experienced benefits of
Tanzania was eaten mainly as enjera and ugali these stoves which were not related only to
after milling, or in dishes which involved cooking time and fuel conservation. They were
pounding and then boiling of the grain. The the benefits of increased safety in an enclosed
level of refinement in the milling process varied. space, and the reduced need for tending, giving
Fermentation was part of the process of making more time for other chores or activities which
enjera, and indimano in Tanzania, which was could be undertaken during the cooking time.
consumed by all age groups. Another apparent adaptation of customary
The cooking methods and cooking times of cooking practice in Nepal and Ethiopia was the
blended food varied over the sites, from no preparation of dishes which were not very well
cooking at all (mostly in Nepal) to reportedly liked and had been more rarely prepared in
cooking for as long as ordinary porridge made their home areas, namely khole (Nepal) and ndeta
from other grain (Tanzania). Whenever (Tanzania). They served to eke out the rice and
cooking of blended food was observed in maize ration respectively, but both were treated
Tanzania, however, cooking time did not with considerable disdain by women in
exceed ten minutes. In Nepal blended food was interviews. This was particularly the case with
commonly fried at high temperature for a few khole, which had been rarely eaten before, was
minutes. The different cooking methods associated with hardship, and was particularly
reflected the way in which similar foods had distasteful when made from the parboiled rice
been prepared formerly, that is in the form of in the ration. Ndete was more acceptable, but
thin porridges made from grain. In addition, had previously been eaten more in some
blended food was versatile, in that it was made regions of the home country than others,
like apuwa, or added to sugared tea in Nepal, or whereas it was now being eaten quite generally.
made to a runny consistency like a milk in It had been more commonly prepared in the
Ethiopia; in other words, it was made to early days of the camp, before mills were
resemble familiar foods. Only in Nepal had the established, but was still being cooked to make
refugees been made aware that blended food the maize go farther and avoid milling costs.
was pre-cooked and that cooking time could be However, women said it could be made only
adapted accordingly. from the white local grain, not the yellow
imported grain.
Cooking methods: There was little suggestion
that refugees had changed any of their methods It is important to note that, in all sites at the
of preparation and cooking, apart from the time of the field study, no problem related to the
awareness in Nepal of the campaign regarding quantity and quality of water was reported to
rice-preparation. What was different were the have any effect on the use of food. Water quality
limitations on the range of methods imposed by in particular is relevant to the cooking of
camp conditions. For example, in Tanzania blended food which, because it requires little to
some women complained about a lack of no cooking, might be mixed with unboiled or
cooking pots and specialised containers for the inadequately boiled contaminated water prior
preparation of alcohol and some specialised to consumption. However, in all sites, quality at
dishes like buswage. In Nepal, there was not source was good, and water-distribution points
enough space inside the hut to make the were within easy reach of the dwellings, so that
preserves, gundruk and sinky, which would significant water storage in the home was not
normally be available in all households to necessary.
accompany main dishes.
Shortage of fuel was common in all sites, and Intra-household distribution
so the speed of cooking was a consideration. Details of intra-household food distribution are
Already-familiar methods of speeding up difficult to ascertain and interpret in most
cooking were being employed, such as mashing settings, and any such studies can produce appar-
down beans and lentils by hand during the ently conflicting information.13 For example, in
cooking process (Nepal, Tanzania), and using Nepal, a field-based quantitative intake study in
soda bicarbonate or ash (traditional soda) to one of the camps had indicated preferential
soften beans and cassava leaves (Tanzania). allocation to males (Baker 1995); a study of
What was new to some was the use of a fuel other household allocation in rural Nepal indicated

Acceptability and use of cereal-based foods in refugee camps

preferential allocation to adult men but not to two days. When blended food was used in this
boys (Gittelsohn 1989); an anthropological way, its micronutrient benefits had not been
study in Bhutan in recent years had suggested maximised (because it was not being consumed
there was no nutritional discrimination between daily), and it had been used inappropriately,
boys and girls (Wikan 1990). because the BFP had been targeted to children.
In the study, in all sites, a subtle bias in alloca- However, it had been used appropriately in the
tion of more or better food to adult men was context of the rationale of the 1996 Food Assess-
suspected by the researcher. For example, in ment Mission (FAM) by WFP and UNHCR, used
Nepal women in interview explained how the to introduce blended food into the general ration.
woman serves members of the household first Breastfeeding was virtually universal in all
and eats her share only when they are satisfied. sites up to at least two years (or next pregnancy
In Ethiopia also, women described the allo- in Ethiopia). Blended food was used as a
cation of food by the woman, who serves herself complementary food in all sites, used from the
only after the men have been served; and in age of four months upwards.
Tanzania the men eat from a separate plate, Thus all members of the household were
served by the woman. eating blended food (mostly in the mornings in
However, women may have presented the Nepal and Tanzania), but children tended to be
interviewer with a 'traditional' representation, given an extra serving later in the day if they
rather than a description of contemporary asked for it and if there were no leftovers or
practice. It was suggested by some informants such-like to give as 'snacks'. However, this is not
that practice differs between the poor and the at all the same as saying that 'blended food was
better-off (Ethiopia), and between the educated eaten mostly by children', an allegation often
and non-educated (Tanzania), and that in the loosely made about blended food generally.
constrained camp situation all have to In addition to food consumption within the
understand the need to share equally the little household, it is worth noting that food was also
they have (Nepal). There were also references consumed outside the home in all sites. For
to camp life and its effects on meal times, with example, in all sites some men were working
different members of the household coming outside the camp on a casual basis and were
and going at different times (Nepal), and the being paid with a meal, or would buy cooked
need to cater for children going to school in food in small restaurants; this food would be
shifts (Nepal and Tanzania). more likely to include meat than the meal
In Nepal and Tanzania, blended food was available at the household. In Ethiopia, many
eaten by all in the household in the morning as men were buying and chewing the stimulant
a breakfast food. Its availability allowed women qaad, a green leaf presumably rich in vitamin C.
more flexibility in the timing of subsequent Women also worked outside the camp, for a
main meals during the day. For example, in daily wage (Nepal), as petty traders (Ethiopia)
Nepal lunch and supper would be earlier in the and selling beer, a few trading farm labour for
day if blended food had not been taken in the food (Tanzania). In Tanzania, schoolchildren
morning. In all sites, it was used as a snack food were receiving fortified biscuits at school.
later in the day for children particularly, instead
of leftovers, which were also commonly used for
this purpose. What made food 'acceptable'?
Blended food was not used as a substitute for
a main meal, except in Ethiopia in the poorer From the investigation into preference, observa-
camp, where women said that blended food was tion of use, and discussion with the refugees, it
taken as lunch or supper when the other meal seemed that there was a range of factors which
that day had been sufficiently 'heavy'. In such influenced the level of acceptability of food
cases, a breakfast of enjera had been eaten. items. Some were of more importance in one
Where it was taken as lunch or supper, the site than in another, and some were interlinked,
monthly ration was finished in one week. Much so that the order in which they are discussed
of the blended food available in Ethiopia was below does not imply order of significance.
not from the general ration, but pre-mix from
the blanket feeding programme (BFP) for Familiarity
children under the age of five, given weekly, Here 'familiarity' is used to imply a recognition of
which was shared by all and lasted only one or and a developed taste for the food. It is thus

Main report

reasonable to assume that familiarity means agency key informant, was more 'psychological'
acceptance and that the converse also applies, i.e. than 'real', in that the allegation of smell was
that the unfamiliar is likely to be less acceptable. linked to general dislike; as unfamiliarity and
However, this was not always supported by the dislike lessened, so did the smell.
field-study findings. For example, yellow maize The different qualities of the types of blended
meal and grain was not familiar to many of the food distributed were an issue in Nepal and
refugees in Tanzania (those who had not already Tanzania. Only one type of blended food had
spent time in refugee camps in Zaire), but it was been supplied in the general ration in Ethiopia,
extremely popular among the refugees and that only infrequently since the beginning
(especially the meal). This was not only for of the year. In Nepal the locally produced
functional reasons: that, being meal, it required blended food was said to be inferior to the WSB,
no milling. It was accorded positive attributes being coarser and smelling of maize, a grain not
such as 'healthy' and 'sweet'. The refugees in highly valued. In Tanzania also the regionally
Ethiopia had had many years to become familiar produced blended food was considered of inferior
with the wheat grain, which had not been a staple quality, again largely because of the coarser
for them previously, and a thriving market had texture. The quality of the beans in Tanzania
developed around it. In Nepal, the unfamiliar was less of an issue than the quantity, except for
parboiled rice had been well accepted within a the one-off consignment of very old beans which
few months. However, it may be significant that, had been distributed earlier in the year and had
although the yellow grain and meal, the wheat caused considerable anger among recipients.
grain, and the parboiled rice had not been Linked with the issue of quality is a less
familiar initially, they were nevertheless easily tangible consideration of the perceived status of
recognisable as a familiar food type. the food, and respect for it. Damaged or old
The concern about familiarity is often raised food, such as the beans in Tanzania and wheat
in relation to blended food: that its novelty and in the repatriation package (Ethiopia), appeared
unfamiliarity may lead to non-acceptance and to cause resentment and anger among refugees,
misuse. In all sites, this did not seem to have not just because of the poor quality but because
been an issue. It was certainly recognisable as a it was demeaning to be expected to accept and
'porridge' (Ethiopia and Tanzania) and as consume such inferior goods. In a similar way,
something like satho in Nepal. It was also then though not creating the same level of feeling,
fashioned into and likened to something familiar, the lentils (Tanzania) were not respected as a
such aspuwa (Nepal) and milk (Ethiopia). food of equal status with beans; and the smell of
maize, not a highly valued grain, was a negative
Quality and status quality attributed to the inferior, locally
An assumption implicit in the original study produced blended food (Nepal). Another
protocol had been that the issue of acceptability interpretation (by a key informant) of this
would relate to blended food more than to the attitude to the locally produced blended food
cereal, at least in situations where that cereal was (Nepal) was that any commodity produced and
the one normally consumed. However, the packaged locally would be valued as inferior to
cereal staples supplied were well differentiated the imported, foreign product, purely because
by the refugees, so that in Nepal and Tanzania it was local.
they recognised different varieties of the staple
cereal, and in all sites different levels of quality Understanding of health and nutrition
were recognised. Different qualities of wheat issues
grain and maize grain had been supplied in In Nepal, after the introduction of blended food
Ethiopia and Tanzania respectively. In Nepal and parboiled rice, there was a clear relation-
polished rice, available in the local market, was ship between levels of acceptability and the
priced according to type and quality. The scope and quality of public information on
parboiled rice supplied in the ration was perceived health and nutrition which was disseminated at
to have been of inferior quality when first intro- that time and subsequently. This was directly
duced and to have improved more recently. A related to the outbreak of beriberi which had
major feature of the rice was the smell of it, affected many people in all the camps. Fear of
which many said had initially been very strong this condition was an incentive to comply with
and off-putting, but which was now reported to measures taken to control it. Parboiled rice,
be much less so. This change, speculated one more than blended food, was associated with

Acceptability and use of cereal-based foods in refugee camps

the decline in beriberi. In Ethiopia and Nepal, traditional classifications of 'hot' and
Tanzania women did not associate blended food 'cold' food were not significant in determining
with any one condition. levels of acceptability, partly because most of the
In Ethiopia, for the previous five years, food items are not prominent in that classi-
blended food had been available only through ficatory system, and also because adherence to
the supplementary feeding programme (SFP), that tradition was common only among the
targeted at children. It is therefore reasonable older people (according to a refugee key
to surmise that the blended food was eaten by all informant). In Nepal, some refugee informants
members of the household in Ethiopia, not were keen not to overemphasise practices or
because it was understood to be the vehicle for a attitudes among the refugees which might be
daily intake of essential micronutrients, but construed by an outsider as not 'progressive' or
because they had always tended to share the 'democratic'. This was probably because there
SFP rations among members of the household. are a number of human-rights groups and
Common to all sites was identification of the political organisations within the refugee com-
blended food with a nutritious porridge which munity who advocate democratic rights and
they had known before, and which in Nepal and reforms for Bhutan.
Tanzania had been promoted through health Other distinctions were made by refugee and
centres in the home countries as 'nutritious' and non-refugee informants to highlight different
'healthy', i.e. pans pitho and musulaki respec- attitudes to 'traditional' or customary practices
tively. In Ethiopia, women described two types among, for example, the 'educated' and 'unedu-
of porridge which they had been used to cated' (Tanzania), the older and younger
prepare for newly delivered women and young generations (Tanzania), pastoralists and town
children, and as a fasting food. dwellers (Ethiopia), civil servants and fanners,
In Nepal and Tanzania, when asked, women high caste and low caste (Nepal). The attitudes
who had received blended food for themselves of the young, compared with those of the older
or for their children through the SFPs insisted generation, were often alluded to, and this was
that they stored and used this supply separately particularly noted by key informants in Nepal in
from that (unsweetened) received in the general relation to the almost universal primary and
ration. In Ethiopia this was less clear, as up to secondary education which was available to
the time of the study women had mostly refugee children and young people, male and
received blended food through the BFP and female, where all groups and castes were mixing
SFP, and only twice through the general ration. intensively on a daily basis. There was also a
Only in Nepal had information been high level of activity promoting adult female
disseminated about intra-household distribution literacy. Nevertheless, some women explained
and optimal cooking methods of blended food. that, while customary behaviour was changing
Women in Ethiopia and Tanzania particularly in the public domain, in schools and women's
claimed they did not know what the ingredients groups and within the sections, traditional
were, nor did they know that it was fortified and observances survived in the private domain of
pre-cooked. the home, including those related to food and
the 'hearth'.
'Cultural' sanctions
Assumptions about acceptability are often 'Resale' value
related to 'cultural' traditions, including taboos. This was undoubtedly a factor in the level of
There was little evidence of such traditions acceptance of the staple grain, the commodity
influencing acceptability of the ration food which was given in greatest quantities, although
items in the three sites. The clearest illustration less so in Nepal, where rice sales were relatively
was in Nepal, where there were caste-related low and rice was overwhelmingly the preferred
rules and observances connected with the 'hearth'. staple among all refugees. In Ethiopia the wheat-
However, the parboiled rice and blended food, grain market-price had held up and was equal
both pre-cooked and so potentially unaccept- to and seasonally higher than sorghum and
able to those of the Brahmin caste, had been maize, and the regional market was thriving. In
accepted by them. They explained that they had Tanzania the white maize grain was in demand
no alternative: they had not chosen to become locally and regionally and had kept its price
refugees and nor could they know who had well. This did not mean, however, that women
been involved in the cooking process. Also in unequivocally wanted these commodities to the

Main report

exclusion of all alternatives. In Ethiopia some refugees did not have the option in any case to
women said they liked to receive sorghum, and exchange parboiled at a favourable rate for
women in Tanzania liked a combination of grain polished rice. If there had been a favourable
and meal. Thus resale value per se was not the rate of exchange, then the pattern of
over-riding consideration. The resale value of acceptability might have been different.
blended food was relatively low, there was not a Another pragmatic consideration was the
significant local market interest, and the vast time taken to prepare and cook the food, but in
majority of households were not selling it. fact this was not mentioned, except where the
Nevertheless, the general level of acceptability was time was longer than it should have been (as
high. To summarise, acceptability of an item could with the old beans in Ethiopia) and to some
mean that refugees liked it to eat and to sell, as extent in the case of dual in Nepal, where people
in the case of the white maize grain and the wheat, bemoaned the cooking time along with the
but it could also mean that they liked it almost monotony of the food.
exclusively to eat, such as the blended food. Thus acceptability was influenced by a wide
range of factors, intrinsic and extrinsic to the
Substitution refugees, in a dynamic relationship. What was
Acceptability of any food commodity generally, clear from the many interviews was that familiar
and blended food in particular, was affected by foods and cooking methods were largely
whether the refugees perceived it to be in the maintained, but not to the exclusion of inno-
food basket at the expense of the reduction or vations which were understood to be beneficial,
removal of another commodity. For example, in such as parboiled rice, blended food, yellow
Nepal, an acceptability study of blended food in maize meal, and improved stoves. The
1994 had found that refugees were prepared to beneficial effect had been clearly demonstrable
accept it, as long as its inclusion did not mean in Nepal, with parboiled rice and the incidence
any diminution of the rice ration (UNHCR of beriberi, and in Tanzania with improved
1994). In Ethiopia the Food Assessment Mission stoves, but was less so with blended food and
(FAM) of 1996 changed the general ration yellow maize meal in Tanzania — but still there
composition by reducing the cereal and oil was a belief in their healthy properties. There is
rations and introducing sugar and blended a case for encouraging better understanding of
food, and in addition a BFP. However, since blended food (including the facts that it is pre-
implementation began in January 1997, sugar cooked and fortified) at distribution points and
had not been distributed at all, and blended through the health and nutrition services, to
food had been distributed infrequently. Distrib- ensure the fullest maximisation of its protective
utions had been monthly but not timely, and potential.
public information and monitoring very poor,
so it was difficult to assess how much refugees
had perceived any kind of 'trade off over the Cereal fortification at the
last year and, if they had, to gauge how much regional, camp, and household
the general dissatisfaction with the amounts and
composition of the food basket had affected the
acceptability of blended food in particular. A report (Beaton 1995) on the fortification of
foods for refugee feeding explored three
Pragmatism distinct approaches to fortification for the
Refugees in all sites, when discussing aspects of general refugee population:
the ration, often complained about inadequate • centralised/regional fortification of all of the
quantity and monotony. At the same time, they staple cereal;
admitted that they had come to accept what was • community-level fortification of all cereal
given, because they had no choice. In the case of before distribution to refugees;
parboiled rice in Nepal, not only had they been • household-level fortification of cereal after
obliged to accept it, but some refugees said they distribution, at the time when 'households'
had 'had to like it', and had in fact developed a grind the cereal or otherwise prepare it for
preference for it over time. It was observed that consumption. (A variant on this is a situation
there was no local market for parboiled rice, in which the household arranges to have its
because it is not commonly eaten by local allocation of cereal ground in a community-
Nepalis, who prefer polished rice. Therefore based small mill.)

Acceptability and use of cereal-based foods in refugee camps

These three broad scenarios are considered In Tanzania, in the neighbouring region of
here in the light of information obtained during Shinyanga, there was the major cargo logistics
the field study. centre at Isaka, for the WFP 'Great Lakes
Regional Cluster'. Here in 1994 UNHCR/WFP
Regional-level fortification i.e. prior to had supported a parastatal milling company
distribution to refugees (NMC) with machinery and setting-up costs to
There was no regional-level milling capacity in mill locally purchased grain for the refugee
Ethiopia. Outside the region, the nearest large camps in Tanzania. The mills supplied were
mill with substantial capacity still had insuffi- prototypes, designed to UNHCR specifications,
cient capacity to meet the potential milling and together with the generators were
requirements of WFP.14 There were other large transported direct from the UK. By the time of
mills in the capital, but use of these would this study, the contract between WFP and the
increase transport costs enormously, as the NMC had been discontinued. The mills had not
route from the sea port to the camps did not operated to capacity since being installed, and
pass through the capital. Another constraint output had declined to less than one-third of
was the supply of grain. In recent months, due potential. The mills had sustained damage due
to donor commitments, WFP had had to borrow to poor maintenance and lack of operational
food internally to meet the demands of the expertise. At the time of this study, WFP was
refugee programme, and function in a 'hand to planning to rehabilitate the machines and re-
mouth' fashion. There was no buffer stock in the install two, to be managed and operated by WFP
country, and food went straight from the port to Isaka. The expected output of the fully
the distribution points. Regional-level or operational set-up was a maximum of 6,000
national-level milling and fortification would MT/month, i.e. more than the current
requirements of the camps in western Tanzania
involve off-loading, de-bagging, monitoring of
(c.3,192MT @ 350gm ppd), of which the study
the process, re-bagging and re-loading,
site is one. The proposed set-up could be
increasing costs of internal transportation, and
modified at modest relative cost to include
insupportable delays. This would not be
fortification (Cawkwell, personal communication).
appropriate in the context of serious resource
Thus the main constraints on fortification here
constraints arid uncertain commitments on the were not primarily technical or exclusively
part of donors. financial. Rather, one issue to consider is how
In Tanzania, regional commercial mills were much the USA, as the major donor to the
being used by WFP to mill small amounts of programme, will continue to send increasingly
grain for the camps. They were of very small larger proportions of maize to the region as
capacity compared with overall requirements. meal, not as grain. Where this meal is supplied
Again, fortification might be theoretically (mostly unfortified, a small proportion partly
possible at these sites, but present output would fortified), then the vehicle for fortification will
be unlikely tojustify the investment. In addition still need to be a pre-fortified blended food.
to output, there were problems of maintenance
Regional-level cereal fortification, i.e.
and subsequent breakdowns. These were
fortification before distribution to refugees,
compounded by the problems of power supply therefore would require the following conditions:
in the city, where rationing of electricity was
increasingly common, owing to the national • adequate milling capacity in a location
effects of drought on the hydro-electric reasonably close to the refugee population,
generating plants. Diesel-powered back-up to minimise extra costs of insurance,
generators were not allowed in mills situated in transport, shipping, and handling;
residential areas. Another problem experienced • a medium- to long-term commitment by
with these mills had been price agreements donors, and a reliable food pipeline;
between WFP and the mill owners, so that • considerable technical and management
expertise to ensure logistical efficiency and
earlier contracts had not been renewed.
quality control at the milling site.
Furthermore, the product of these mills was
found to have been relatively unacceptable to It would be most appropriate where the milled
the refugees. Thus issues of scale and quality of cereal was highly acceptable and so was mostly
output, reliability, and continuity would need to eaten, not sold or exchanged. This strategy
be addressed in considering investment in could be appropriate in an emergency phase,
fortification at this level. where the equipment and expertise was already

Main re poll

in situ, and particularly appropriate where • training and monitoring of mill technicians,
milling capacity was inadequate at camp level. who would be the ones responsible for
measuring and mixing the fortification mix.
Camp-level fortification, i.e. post-
distribution to refugees It would be most appropriate where the staple
grain was eaten in milled form and highly
There were mills in the camps in Ethiopia and acceptable; or, if not, was sold or exchanged to
Tanzania. Their capacity was around 1MT/12 obtain an alternative whole grain eaten in
hours (single-piston engine) and 3MT/12 hours milled form.
(twin-piston). In Tanzania, the mills were all This would be an inappropriate intervention
privately owned and the machines and spare in an 'emergency' phase of a refugee situation,
parts were available locally and were of the same where the infrastructure, including milling
type as the local village mills. The total capacity
capacity, had not yet become well established.
of milling in Tanzania could be matched with
the amount of grain distributed, assuming a 24-
hour day and all machines working to full Household-level fortification, i.e. post-
capacity; but it is unlikely that technically an distribution to refugees
automated feeder could be attached to a mill of
this size. If it could, or if manual mixing was Fortification of cereal at the time of grinding or
considered, trained personnel would be necessary pounding at household level did not appear to
to monitor the feeding mechanism to ensure be a feasible option, for the following reasons:
correct mixing levels. The extra costs would • It would be difficult to ensure accurate
have to be borne by the implementing agency measurement and thorough and even
and not the refugees, as milling costs already mixing, given the limitations on time and
restrict the amount of grain which is milled. space, and the absence of standard
However, if technical and economic problems measures.
were resolved, another consideration must be • When not sold or exchanged, the staple
that a proportion of the maize grain in Tanzania grain was not always milled for
was sold to obtain other food commodities consumption, as in Ethiopia and Tanzania,
which did not require milling; and/or the grain where dishes were commonly prepared
was eaten after pounding and cooking as whole which required some pounding or no
grain. This would mean that the investment in pounding.
fortification at camp level might not result in • Where the staple grain was milled, it was
consistently higher levels of micronutrient mostly done at the commercial machine mill.
intake of all refugees via the ration grain. However, a household-level fortification mix
A further limitation on camp-level could be added as a seasoning during the
fortification arises from the situation described cooking process. In each site, some seasoning
in Ethiopia. Some of the mills were privately was added to the relish towards the end of the
owned; others were owned by WFP, and cooking time. In Nepal a pinch of turmeric was
managed by the implementing partner. These always added to the vegetable curry; in Ethiopia
were found to be functioning far below capacity, salt and spices were added to some dishes; in
and the reasons for this were not clear. The Tanzania salt only was added. The turmeric was
reasons were not related to technical matters, to enhance the colour and appearance, the salt
but to management constraints and probably to and spices to enhance taste. It is assumed that a
competing commercial/political interests. These fortification mix would have no colour-
matters would need to be resolved before enhancing or taste-enhancing properties, so the
investing in fortification at these facilities, and amount added would need to be pre-measured
the relationship clarified between the com- and adjusted according to how much food was
mercial and free mills at the same site. being prepared at the time. However, if the
Thus camp-level fortification would require fortification mix was in a carrier which was
the following conditions: flavoured, like a salt, then the amounts could be
• adequate milling capacity for the milling regulated by taste, not by measurement.
needs of the whole population; Alternatively, the mix could be added to the
• affordable and accessible mills, so that food as it is served. This would be an option for
people were not milling large amounts in example in Nepal, where food is served on to
their houses; each individual's plate; but in Ethiopia and

Acceptability and use of cereal-based foods in refugee camps

Tanzania, groups within the household share agreed with WFP, so that the product had been
from a single plate, and, once served on to that unavailable for distribution for three of the five
plate, the mix could not be distributed evenly. monthly distributions of 1997, up to the time of
The mix would have to be in the form of an this study. There was now at least one other pro-
additive which could be sprinkled to taste in the ducer of blended food, operating on a smaller
food while cooking or when on the plate. There scale as yet, but WFP was already a customer.
might be sensitivities and suspicions to be In Tanzania, a blended food produced in
overcome which would not be apparent in neighbouring Kenya was procured by WFP for
interview (in Nepal only, women were asked a the refugee programme recently, when there
hypothetical question about adding a new were problems in the CSB pipeline. However,
substance to their food, and they said that they there had been serious problems with the
would comply), but which would emerge only in quality of this consignment, and a follow-up
field testing. investigation was being undertaken by WFP at
the time of the study.
Production and fortification of blended While all producers were required by WFP to
food use a specific imported vitamin/mineral pre-
The other fortification strategy is the provision mix in fortification of the blended food ordered
of a pre-cooked and fortified blended food by WFP, there was no facility in-country to
which is given in addition to the staple cereal. analyse samples for their micronutrient
This is distinct from the strategy of cereal content, so a key aspect of quality control
fortification discussed above. remained problematic.
For a summary of the properties of the Blended food in ration, i.e. fortification
blended foods in use in the three study sites, see before distribution to refugees, requires the
Appendix l(d). Locally produced fortified following factors:
blended food had been distributed in the • small-scale milling infrastructure;
general ration in all sites (in Tanzania the • quality-control mechanisms.
blended food was produced in a neighbouring
country, not in the host country). In Nepal and It is most appropriate in the emergency phase
Tanzania, imported blended food (WSB and of refugee situations, where a milling
infrastructure is not in situ, and in protracted
CSB respectively) had also been distributed
phases of refugee situations, where diversifica-
interchangeably with the local product. In both
tion of the diet and cereal fortification are not
cases, there was a preference among the
viable options. Even where these latter conditions
refugees for the imported product, but the local
do not pertain, blended food constitutes an
product had not been rejected. The reason
excellent complementary food for infants.
given for this preference was quality of the
product; in both cases this referred to texture, There tends to be a generally high accepta-
and in Nepal also to smell. bility among all age groups. Acceptability could
be reinforced if sugar were provided and
In Nepal, the main problem with the locally information disseminated about nutrient content.
produced product was the price, which was Use could be improved with cooking infor-
higher than that of the imported WSB. The mation and provision of containers for storage.
manufacturer said the price could be reduced if
consistently large orders were placed by WFP,
i.e. if demand was assured. WFP could not do Conclusions
so, because of the time limits of its operational
budgets. Also, if donors offered WSB to WFP, The case-studies illustrated a wide range of
the latter had to accept the donation and so factors which influenced food use and accepta-
could not order the local product too far in bility. These included level of donor commit-
advance, in case of duplication. ment, relationship with markets, regional and
In Ethiopia, the factory which has produced local, distribution systems, and the quality of
blended food for many years was undergoing health and nutrition services. Those important
major managerial changes and was operating factors more directly related to the commodities
now as a profit-making company, supported no included familiarity, quality, and resale value.
longer with subsidies and tax exemptions. The Blended food was used and liked widely, in
factory was operating far below capacity and all sites and by all members of the household. It
had been failing to meet the delivery schedule was not used only for children, or perceived as a

Main report

food for sick children only, although in all sites, for regional-level milling and fortification at a
while adults ate it once a day, children when site which is already a major logistics base.
possible ate it twice or more. There were However, the capacity of the local town mills
preferences for the USA product rather than was inadequate for milling and fortification pre-
the local, and for sugar to be added. A common distribution. The capacity of the Muyovosi camp
reason for not selling was that they were given mills could be adequate for milling post-distrib-
too little. However, neither the local product ution, but operational issues of fortification at
nor the unsugared products were rejected on this level would have to be fully researched. In
that basis. Nor does it follow that increasing the Ethiopia, there were severe resource constraints
amount would lead to more being available on which meant that milling and fortification in
the market, since it was perceived to be a healthy country was not an option at the time of this study.
food, for which there was no available substitute In summary, blended food seemed to be well
which would be so easy to prepare. It was accepted and widely consumed. Appropriate
acceptable on grounds of familiarity, because it use might be enhanced by increased provision
was recognised as a porridge mixture, and also of public information about its specific qualities.
was easily made into a familiar porridge-type Acceptability was affected by operational issues
dish, or other dishes which were familiar. It was such as the quality of the local product, rather
thus a versatile asset. It was particularly valued than the product^m'. Along with technical and
where its properties were known and it was operational considerations, the cereal-forti-
understood to be directly related to improved fication strategy option involved aspects of
health status. Use (to a small extent) and accep- acceptability, such as the type and quality of the
tability (to a larger extent) were influenced by milled cereal and its resale value, and aspects of
the perceived quality of the locally produced use, such as whether and how the cereal was
product. A technical and operational issue for milled prior to consumption.
donors therefore (but not related to use and
The studies illustrated that refugees did not
acceptability) is the quality control and monitor-
accept and use food purely on the basis of its
ing of the micron utrient. content of the local
nutritional content. The value given to it was
not dictated by any one factor such as
Use of the cereal food was more diverse than nutritional content, ease of preparation, market
that of blended food, and at different levels: price, or familiarity. The balance of influencing
extra-household, inter-household, and intra- factors shifted with circumstances, with range of
household. Uses ranged from its being mostly opportunity, and with new and established
eaten (Nepal), being mostly eaten and sold/ preferences. Refugees in these case-studies thus
exchanged for alternative staples (Tanzania), to appeared to be more flexible, adaptable, and
being eaten and also sold in bulk (Ethiopia). pragmatic than is often assumed.
There were also a number of cooking methods
which did not always require milled grain.
Where it was required, there was little hand- Notes
grinding in the African sites (not applicable in
Nepal), with commercial machine mills being 1 For definition and composition, see section
widely used, and dishes being prepared with 'Cereal fortification and blended food', on
dehusked (pounded) unmilled grain. page 13.
Even within the same cereal type, there were 2 'Responding to the Nutrition Crisis among
degrees of acceptability, which were related less Refugees: the Need for New Approaches',
to familiarity and more to perceived quality and under the auspices of the Refugee Studies
resale value. Where meal was given, for example in Programme, Oxford University, March 1991;
Tanzania, it was preferred to grain and not sold, and the 'Machakos Workshop on the Imp-
except when it was the regionally milled meal, rovement of the Nutrition of Refugees and
which was widely considered unacceptable. Displaced People in Africa, December 1994'.
Thus while the issue of in-country fortification 3 For a wider review of outbreaks, see the
remains primarily a technical and economic one Reports on the Nutrition Situation of Refugees
for donors, these situation-specific aspects of and Displaced Populations, published by the
cereal use and acceptability would also have to ACC/SCN.
be taken into account in reviewing options. 4 'Acceptability' of a food in this study is defined
There was considerable potential in Tanzania as the degree to which refugees expressed a

Acceptability and use of cereal-based foods in refugee camps

liking for the food, and is assessed according 9 The refugees referred to in this table are
to scientific parameters such as organoleptic those arriving in 1988, not subsequent influxes
properties (like taste and texture). of 199 land 1994.
Studies which have looked at similar aspects 10 Another problem of local supply, rather than
include one among refugees from Myanmar access, and peculiar to Ethiopia was a serious
in Bangladesh, where an Instant Blended shortfall in the supply of blended food. This had
Food product for supplementary feeding was been re-introduced into the ration at the begin-
found to be acceptable, adaptable, and valued ning of 1997, and a supply ordered locally by
(was not sold) (Gladwin 1993 unpublished), WFP. However, serious management problems
and one examining use of ration items among at the manufacturer's meant that the
refugees in Latin America (Hansch 1992),. required amounts had failed to arrive in time.
In the text, 'milling' refers to milling whole 11 The smell may be determined by the
grain to meal, either by machine or by hand standards of processing, so that if the water is
grinding. It is not used to signify the first changed between batches, the smell will be
stage of the milling process, which is removal less than if repeated batches are boiled in the
of the inedible outer layer of the grain (by same unchanged water.
hand this is done by pounding to whole 12 The reasons for the continued reporting of
grain). Fortification pre-mix cannot be added vitamin B deficiency, albeit at low prevalence
to whole grain, only to meal (flour). rates, are not clear and are discussed in the
Therefore it is extremely difficult to fortify Nepal case-study.
whole rice grains, unless they are milled to 13 Re. intra-family household distribution in
flour. Instead, parboiling slightly cooks the South Asia, 'quantitative and gender aspects
rice, makes the outside of the grain harder, of allocation are so under-researched by
and conserves nutrients in the grain. anthropologists, and the small body of nutri-
The term 'acute malnutrition' in the text is tional evidence is presented in such a timeless
defined according to ACC/SCN RNIS and classless way by nutritionists, that state-
indicators of wasting, i.e. less than -2SDs or ments of general principle are made at one's
sometimes 80% wt/ht by WHO/NCHS/CDC peril. Caution is also needed over the discrep-
international reference values, usually in ancies between what people say and believe,
children of 6-59 months. Includes severe what they actually do, and what they think
wasting, i.e. less than -3SDs or about 70% about these discrepancies' (Harriss 1990:353).
wt/ht, and oedema, being the key clinical sign 14 This was the Dire Dawa mill, and estimates of
ofkwashiorkor. its capacity ranged widely from c.300 to
See Table 2. 1,000MT/month (WFP key informant).

The top line of the left-hand column
on this page should read as follows:

liking for the food, and is not assessed according

Appendix l(a): Methods

Background • Refugee population as big or bigger than
local neighbouring population, who are
unable or unwilling to offer support.
Research team
In addition to this, the selected site had to be
The field research was led by a single researcher reasonably stable and secure, thus not placing
(Catherine Mears), who worked with a different the research team at risk.
female interpreter (non-refugee) in each of the
three case-study locations. No other field
assistant was employed. Interpreters were
chosen partly for their ability to translate and
Key-informant interviews and
partly for their lack of prior connections with observation sessions
the refugee population and the relief
programme — which would ensure that they Key informants were chosen according to the
had few preconceptions about and affiliations following criteria:
with the group under study. In Nepal the • Senior programme officers of operational
interpreter was a local Brahmin woman; in UN and NGO agencies, in the capital and at
Ethiopia a Kenyan Somali who spoke English, regional and camp levels.
Somali, and Swahili was chosen; and in • Technical officers of UN and NGO agencies
Tanzania the interpreter was a Kirundi speaker. in the health, nutrition, and community-
The researcher and interpreter introduced services sectors in the capital, and at regional
themselves to agency personnel and refugees as and camp levels.
employees of Oxfam UK and Ireland, a non- • In the camps, representatives of
government organisation which was known in management, health and nutrition staff, and
all three sites as an operational agency. other workers, for example in community
services, education, adult literacy.
Local introductions • Refugee representatives, particularly
Contact was made with the relevant govern- women's representatives, and other refugees
ment officials, UN staff, and NGO representa- with particular knowledge, for example a
tives in the capital or regional centre, and with Bhutanese refugee nutritionist in Nepal who
the local administrative centre responsible for had formerly worked at Ministry level.
the refugee camp. A research protocol was pre- • Factory managers, mill owners at camp and
pared in advance of the studies, and distributed local levels, a milling consultant (Tanzania).
widely by the researcher in each field site. The Other informants were interviewed on an informal
researcher then personally explained the basis, for example, refugees working with agencies
research objectives, scope, and type of study. in the camp; in the text these are referred to as
'informants' and are not listed in the relevant
Choice of study location Appendix. Informal unplanned interviews and
Before selecting the countries and refugee observations in and around the camp were record-
populations for each case-study, certain ed in the field notebook as soon as possible. All
characteristics were specified that were felt to notes were later written up in full and then coded
typify nutritionally vulnerable refugee according to the study questions in the protocol.
populations (described in the original research Afieldjournal/diary was maintained throughout.
protocol). These included the following: Initially, the idea of partaking in and observ-
• General ration set by UNHCR and WFP at a ing family meals was included in the plan, but
minimum of 1,900 kcal (which assumes this proved not to be feasible, as both the
nutritional dependency). researcher and interpreter were newcomers in
• Blended food included in the ration. the camps, and the schedule did not allow enough
• Limited access by refugees to local markets. time to develop appropriate relationships before
• Few income-earning opportunities for the study began. A major consideration also was
refugees. that the presence of strangers observing might
• Limited availability of or access to cultivable influence what happened during the observa-
land. tion period. In all sites, the researcher was

Acceptability and use of cereal-based foods in refugee camps

inevitably identified by refugees with the donor Data from observation sessions were recorded
and distributing agencies, and this influenced in writing at the time of the observation. The
behaviour and responses in interviews. interpreter was asked to make mental and
However, informal and formal observation of sometimes written notes also for cross-checking
cooking of the staple1 cereals and blended food afterwards. Photographs were taken during
was undertaken. The sessions were arranged casu- formal and informal observations.
ally with only one or two women and their friends,
one or two days in advance. In Nepal, for example,
the researcher asked for women to demonstrate Household-level interviews
cooking in their huts. A demonstration of prepar-
ing a family lunch of rice and vegetable curry in a The household here is defined as, 'A group of
hut was arranged with two days' notice, but demon- people who "share a fire", that is, who share
strations of die preparation of blended food by food on a regular basis. Households usually
adding hot water and by frying (each in individual contain people related by blood or ritual (such
refugee huts) were done on the spot by a woman as marriage) called families. People who are not
member of the Refugee Women's Forum. members of the family may also live in a
In Kebribeyah camp, Ethiopia, two cooking household' (Scrimshaw and Hurtado 1987).
demonstrations were arranged; each one took In the field studies, the household did not
up over half a day. The researcher identified a always correspond to ration-card groupings of
woman who had been responsive during the its members, so that, for example, adult
interview, and then asked her and one or two bachelor sons living in a separate hut might
friends or neighbours if they could cook certain have their own ration card, but eat food
dishes at home the following day. The researcher together with and cooked by their mother.
bought the food the day before and, after In each site, between 20 and 30 households were
preparation and cooking, ate it together with visited in total, and interviews conducted which
the women. All demonstrations appeared to be varied in lengthfrom40 minutes to over two hours.
authentic in terms of showing the daily routine Usually no more than three interviews would be
activity (apart from the curious observers who conducted in one day. Where the main interview-
inevitably gathered from nearby huts). ee (the person responsible for preparation and
Casual observation and informal interviews cooking of food, invariably a woman) was joined
also took place during visits and walking through by members of the family, one or two neigh-
the camps. In addition, formal observation of bours or friends, the interview continued and
food distributions was undertaken in two sites: included their contributions also. This was a positive
• In Nepal, a ration distribution was observed advantage in most cases, as the main interviewee
with no prior arrangement being made. tended to be less reserved and more relaxed.
• In Tanzania, one ration distribution was Selection of households
observed formally and another informally.
The households were chosen by the researcher,
In Ethiopia, the end of a ration distribution was together with the interpreter, and in Ethiopia
observed, not the entire process, as prior notice with guidance from a women's representative
of distribution was not available. and in Tanzania with help from the Oxfam
Markets inside and outside the camps were woman community worker. The selection was
visited, and market prices collected. Mills and fac- made to ensure that visits were made as widely
tories were visited at camp and regional levels: as possible within the refugee group.
• In Nepal, the blended-food factory In Nepal, where 90,000 refugees were
(Mahalaxmi Foods, Biratnagar) was visited situated in seven separate camps, all within a
by prior arrangement through WFP. one-hour radius of the researcher's base, all
• In Ethiopia, camp mills and town grinding camps were visited. The camps were divided
mills were visited on an ad hoc basis. The into sectors and sub-sectors, with ethnic groups
blended-food factories in Addis were visited mixed throughout. Sometimes a household was
by prior arrangement through UNHCR. chosen because cooking was seen to be under
• In Tanzania, mills were visited in the camps way in the house, or in one case because it
and in Kigoma town, and a visit was made to included a very old woman, or twice following
the WFP logistics base in Isaka, Shinyanga up a suspected vitamin-deficiency case described
Region. by the Health Centre. Thirty households were

Appendix I (a)

visited as planned, plus two Nepali households The refugees, when approached and asked if
in villages close to the largest group of camps. they could spare some time to discuss issues of
In Ethiopia, a small camp called Kebribeyah cooking and food, were usually polite and
(KB) of c.l 0,000 was chosen as the main field welcoming. In Nepal, effort was made to
site, as it was distinctly poorer and more approach the huts quietly and unobtrusively
dependent on the general ration than the and slip inside the hut when invited; but the
others and so fitted the study criteria more huts in all Nepali camps are very close together,
closely. It was the only camp where the residents and someone who is obviously a foreigner
are not surrounded by their own clan and immediately attracts attention and inquisitive
therefore they receive less support from the followers, particularly children. The huts are
'local' population. All the main sections of the bamboo and twig lattice, so that a quiet, private
camp were visited. However, for purposes of interview was difficult to achieve, though on
comparison and triangulation, a second camp, some occasions it was possible (especially when
Hartisheikh A (HA), was chosen; this also the children of that section were in school). In
ensured a wider range of interviewees, since Nepal, people tended to speak quite openly,
those in KB were mostly from southern and and fear or extreme reserve were not apparent
central Somalia, although from both pastoralist in most of the interviewees; but, as in all
and urban backgrounds. By the end of the communities, some were much more respon-
study, twenty households had been visited and sive and informative than others.
one household in a local village. Care was taken In contrast, in Ethiopia, key informants advised
to visit a range of geographical sections and a that people in the camps were generally unhappy
range of clans. The researcher and interpreter with visitors continually coming to do research
were accompanied at all times in both sites by a and assessments. The refugees generally seemed
woman from the Women's Committee, without very guarded about disclosing sources of food
whom it would have been difficult to gain access. outside the ration, and income opportunities.
In the Kigoma region of western Tanzania This could be explained in a number of ways,
there are several camps, which accommodate including perhaps a sort of'survey fatigue'among
either Burundian or Congolese refugees. Of refugees, a long history in the camps of changes
these camps, Muyovosi was selected, because it in ration composition and amounts delivered,
most closely corresponded to the original disputes over numbers of beneficiaries, and the
criteria specified for the selection of sites in the present climate of renewed interest by agencies
research protocol. There are 11 zones in in repatriation and all that it implies. For these
Muyovosi camp, so it was decided to interview reasons it was stressed to the refugees that the
two households per zone, making a total of 22 researcher had come from England 'for studies'
interviews. and was not representing UNHCR or WFP.
In all study locations, access to the camps was In Tanzania, if women were shy and unforth-
obtained by arrangement with the refugee- coming, it was discovered that they were more
camp managers, and then on a daily basis with relaxed if friends were invited to join them. It
the practical support of women's repre- also meant that the suspicion of neighbours was
sentatives (Nepal and Ethiopia) and a woman mitigated by having someone else present.
community worker (Tanzania). Sampling was The semi-structured interviews were based
'opportunistic' (Scrimshaw 1990), in that on a simple checklist, which helped to guide the
women were selected 'on the spot' during the interview, where many questions were
visit to that section and if, when asked, they formulated based on the interviewee's response
agreed to give time and space for the interview. and the critical areas of interest for the
The interviews almost always took place interviewer. This meant that, as new issues
inside the woman's house, not in a public space. arose, more information could be gathered in
When a household had been chosen, the subsequent interviews in an iterative process.
researcher asked to interview the person who Certain issues of interest were followed up and
did the cooking. The link with Oxfam was were later deemed more or less relevant than
explained at the outset, and interviewees were expected, over time and during analysis and
told that a few households were being visited feedback sessions. In the main report, reference
just by chance and that no names or hut is made often to 'women in interview', because
numbers would be recorded. (Most seemed the vast majority of interviewees were women.
unconcerned about the last point.) However, on a number of occasions, men were

Acceptability and use of cereal-based foods in refugee camps

present, and occasionally the main interviewee A preliminary case-study, including contextual
(except in Ethiopia) was male; this is indicated in information, was composed during the time of
the list of interviewees attached as an appendix to the field study. Care was taken to minimise the
each case-study. Household interview data were main sources of analytic bias identified by Miles
recorded as fully as possible in thefieldnotebook and Huberman (1994).3 The availability of
at the time of or just after the interview, relevant secondary data differed between study
depending on the attitude of the interviewee. sites, and this resulted in differences in the
scope and depth of contextual material
obtained and included in the case-studies.
Checklist for household 2 Triangulation was done by consulting
semi-structured interview secondary sources as widely as possible, and in
Nepal and Ethiopia visits were made to one or
two local households (not refugees). In each site
Household: Income/occupation now and also, formal feedback sessions were arranged
formerly. Education — formal/non-formal with groups of refugee women at the end of the
period of household visits, and also formal
Types of blended food and cereal feedback sessions with agency representatives.
obtained: From where? By whom? The preliminary draft of the case-study was
How much? How often? Then what? used as a basis for these sessions. These sessions
were invaluable not only for triangulation, but
Storage of blended food and cereal: also for clarifying issues, disseminating
How? Where? Milling? findings, and ensuring transparency.
The case-studies were reviewed with the
Preparation: Who? How measured? research supervisor following each period of
Cleaned? Washed? Other foods mixed? field work, while the external adviser submitted
comments. On completion of all three case-
Cooking: Who? Where? When? How? studies, a matrix (or spreadsheet) was
With what else? developed, showing key findings related to each
of the main research questions in each study
Distribution: By whom? To whom? site. This summary allowed a rapid review of the
When? How often? How much? Degree main features of each case-study, which served
of commensality? Leftovers? as the basis of two days of discussion between the
Exchange/barter/buy/seasonal effects? field researcher and the research supervisor.

Discussion/observation/points of note
1 The main part of most meals; for example,
Data analysis thick porridge made from maize or cassava or
rice. Staples are usually cheap and provide
A diary, field journal, and interview notebooks most of the energy, protein, and fibre in a
were maintained throughout. Field notes taken meal, and some vitamins (Savage King and
at the time of household and key-informant Burgess 1992).
interviews were transcribed within two days and 2 This is not the layout of the original checklist.
arranged according to the study questions, with 3 These are the 'holistic fallacy', where events
the use of 'Procite', a bibliographic database are interpreted as more patterned and
programme (supplied by PBS Europe). congruent than they really are; the 'elite bias',
A glossary was developed in each study site to where data from articulate high-status
clarify some of the most commonly used words. informants are given more weight than those
Definitions are mostly from discussion with from less articulate, lower-status ones; and
interpreters and sometimes other informants; a finally the loss of perspective as the
few are drawn from secondary sources. The researcher is co-opted into the perceptions
glossary is not intended to be comprehensive or and explanations of local informants (Miles
definitive in any way. It is rather a list of words and Huberman 1994: 263).
and their meanings as used by the refugee
participants in the study.

Appendix l(b): Fortified pre-cooked blended food
— definitions and examples

Definitions * vitamin/mineral supplement;
• sugar, if required, in which case it replaces
'Food fortification may be broadly defined as the an equivalent amount of cereal.
process whereby one or more nutrients are added to It is manufactured according to the following
foods to maintain, improve or enhance the quality of a recipe (other ingredients and combinations are
diet consumed by a group, community or population' possible):
(Henry and Seaman 1992).
Whole maize: 80 per cent by weight
Blended food1 should be produced in Whole soya beans: 20 per cent by weight
accordance with the 'Code of Hygienic Practice Vitamin/mineral pre-mix (as specified below)
for Foods for Infants and Children' and 'Code
of Sound Manufacturing Practices' of the Codex It should be manufactured by means of
Alimentarius. It is a mixture of the following extrusion or roasting/milling. It should be
ingredients: fortified to the extent that to each metric tonne
of finished product are added 1 kg vitamin pre-
• cereal such as maize, sorghum, millet, wheat mix and 3kg mineral pre-mix (obtained from La
or a combination, providing carbohydrates Roche Ltd. Switzerland, or its local authorised
and protein; dealer)(WFP, Rome, undated).
• pulses (chickpeas) or soya beans as an
additional source of protein;
• oilseeds (groundnuts, dehulled sunflower
seeds, sesame), soya bean or stabilised
vegetable oil as an additional source of oil; Examples:

Fortified pre-cooked blended foods available in field study sites


Manufacturer Mahalaxmi Protein Grain Faffa factory, HCFM Protein Grain
Foods Biratnagar Products Addis Ababa Addis Ababa Products
International International

Contents wheat pre-cooked wheat pre-cooked maize pre-cooked maize pre-cooked maize (processed,
maize pre-cooked soya flour soya flour full fat soya pre-cooked gelatinised)
soya pre-cooked salad oil sugar chickpea pre-cooked soya flour
vit/min pre-mix vit/min pre-mix vit/min pre-mix sugar (defatted, toasted)
vit/min pre-mix soya oil
vit/min pre-mix

Process roasting extrusion roasting roasting extrusion

Food values 400 Kcal 360 Kcal 402 Kcal 380 Kcal 380 Kcal
per lOOgm 14g protein 20g protein 14.7g protein 13.3g protein 18g protein
dry product 6gfat 6g fat 7gfat 7.4g fat 6g fat
60g carbohydrate 70. lg carbohydrate 65g carbohydrate 60g carbohydrate

Cost Ex-Factory2 ? US$350 US$3903 c. US$450 c.US$450 US$320 3

Preparation None Famix:water 2:5 Tenamix:\vater 2:5
Instructions 5-10 minutes 2 teaspoon oil
boiling Cook for 10 minutes

Acceptability and use of cereal-based foods in refugee camps

Micronutricnt specifications (per lOOg dry finished product)

WFP Rome Unilito WSB Famix Tenamix CSB ex-USA
recommen- ex-USA

Vit. A 1,664.0 i.u. 400.0 1,658 i.u. 1,300.0 i.u. 1,500.0 i.u. 1,700 i.u

Vil.Bl 0.128mg 0.1 mg 1.49mg 0.1 mg 0.3mg 0.7mg

Vit. B2 0.448mg l.Omg 0.59mg 0.4 mg 0.5mg ().5mg

Vit. B3 4.8mg o.Omg 9.1mg 5.0mg - 8.0mg

Folate 60.00 50.00 50.00 0.06mg
microgram microgram

Vit.C 48.0mg 50.0mg 40.0mg 30.0mg 20.0mg 40.0mg

Vit.B12 1.2 5.0 4.0 1.0 0.3 4.0
microgram microgram microgram microgram microgram microgram

Iron 8.0mg 15.0mg 20.8mg 8.0mg 12.0mg 18mg
(as ferrous

Calcium lOO.Omg lOO.Omg 749.0mg lOO.Omg 200.0mg SOO.Omg
(as calcium (? not as (? not as
carbonate) calcium calcium
carbonate) carbonate)

Zinc D.Omg 5.0mg 4.6mg S.Omg 1 O.Omg 3.0mg
(as zinc

Vit. B6 - - 0.52mg - 0.4mg 0.7mg

Iodine - - 50 - 0.05mg 50
microgram microgram

Magnesium - - 202mg - 20.0mg lOOmg

Selenium - - - - 25.0mg -

Potassium - - 624mg - 164.0mg 700 mg

1 The following description is adapted from 2 Excluding costs of insurance, shipping,
WFP Rome: 'Processing Instructions and handling, and transport at time of study.
Product Specifications for Precooked 3 From'Prices, nutritional value and unit cost of
Fortified Blended Food, particularly for WFP-supplied commodities January 1997'
Older Infants, Young Children, Pregnant (WFP Rome).
Women and Nursing Mothers'(undated).

Appendix l(c): Prices, nutritional value and unit cost
of WFP-supplied commodities, January 1997

Nutritional value/1 OOg Cost per Unit (US cents)

(US/t) (Kcal) (g) (g) (lOOOKc) (100g)

Wheat 200 330 12.3 1.5 6.1 16.3
Rice 300 360 7.0 0.5 8.3 42.9
Sorghum/Millet 185 335 11.0 3.0 5.5 16.8
Maize 185 350 10.0 4.0 5.3 18.5
Cereals, General (EMOPs) 210 - - - - -
Maize meal 240 360 9.0 3.5 6.7 26.7
Wheat flour 290 350 11.5 1.5 8.3 25.2
Bulghur wheat 230 350 11.0 1.5 6.6 20.9
Corn soya blend 320 380 18.0 6.0 8.4 17.8
Wheat soya blend 390 370 20.0 6.0 10.5 19.5
Soya-fortified bulghur wheat 250 350 17.0 1.5 7.1 14.7
Soya-fortified maize meal 220 390 13.0 1.5 5.6 16.9
Soya-fortified wheat flour 240 360 16.0 1.3 6.7 15.0
Soya-fortified sorghum grits 200 360 16.0 1.0 5.6 12.5
Dried skim milk (enriched) 2,100 360 36.0 1.0 58.3 58.3
Dried skim milk (plain) 2,000 360 36.0 1.0 55.6 55.8
Dried whole milk 2,500 500 25.0 27.0 50.0 100.0
Canned cheese 2,500 355 22.5 28.0 70.4 111.1
Canned meat 2,300 220 21.0 15.0 104.5 109.5
Dried salted fish 3,100 270 47.0 7.5 114.8 66.0
Stockfish 4,500 - -
Canned fish 2,400 305 22.0 24.0 78.7 109.1
Vegetable oil 900 885 - 100.0 10.2 -
Butter oil 2,100 860 - 98.0 24.4 -
Edible fat 1,100 900 - 100.0 12.2 -
Beans 450 335 20.0 1.2 13.4 22.5
Peas 340 335 22.0 1.4 10.1 15.5
Lentils 365 340 20.0 0.6 10.7 18.3
Sugar 410 400 - - 10.3 -
Dried fruit 1,200 270 4.0 0.5 44.4 -
Dates 1,900 245 2.0 0.5 77.6
Tea (black) 2,000 - - - - -
Iodized salt 210 - - - - -

•Updated by Budget Branch (FSB)

03 01 97
WFP, Rome

Acceptability and use of cereal-based foods in refugee camps

Appendix l(d):
Composition and nutritional analysis of planned rations

Nepal Nepal Ethiopia Ethiopia Tanzania ' Tanzania
'93 '94 '96 '97 '97 '97
Foods in the ration:
Polished rice (g) 450
Parboiled rice(g) 430
Whole wheat grain (g) , 500 400
Maize flour (g) 350'
Whole maize cereal (g) 400
Lentils (g) 60
Pidgeon peas (g) 60
Beans (g) 120' 120
Vegetable oil, fortified (g) 25 25 30 25 20 20
Salt, fortified with iodine (g) 7 7 5 5 5 5
Vegetables' (g) 150J 100
Sugar (g) 20 20
Wheat Soy Blend (g) 40
Famix (g) 0° 30
Corn Soy Blend (g) 30 30
Amount of blended food in None 40g None 30g 30g CSB 30g CSB
ration: WSB Famix
WHO' Percentage of recommended nutritional rec uirements met°
Energy 2100 kcal 99.3% 106.0% 86.4% 79.0% 90.5% 92.9%
Fat (% of energy) 13.0% 13.4% 19.7% 18.9% 16.5% 17.9%
Protein (% of energy) 8.7% 9.0% 12.9% 12.3% 13.6% 13.3%
Protein 46 g 99.4% 109.7% 128.3% 112.3% 142.2% 142.7%
Water soluble vitamins
Thiamine 0.9 mg 95.6% 210.0% 255.6% 220.0% 211.1% 243.3%
Riboflavine 1.4 mg 18.6% 47.9% 35.7% 41.4% 47.1% 79.3%
Niacin 12 mg 37.8% 55.7% 212.5% 189.4% 41.0% 70.0%
Vitamin C 30 mg 26.8% 75.0% 0.0% 32.1% 60.0% 42.9%
B12 0.9 Tg 0.0% 177.8% 0.0% 33.3% 133.3% 133.0%
Folic Acid Total 160 Tg 110.6 130.6% 271.9% 244.3% 135.0% 102.5%
Fat soluble vitamins
Retinol equivalents total 500 Tg 46.7% 85.2% 57.3% 71.1% 67.0% 66.6%
Vitamin O 3.8 Tg 0.0% 750.0% 0.0% 0.0% 562.5% 562.5%
Iodine 150 Tg 147.0 160.0% 103.3% 110.7% 113.0% 129.3%
Iron 20.4 mg' 33.0% 76.3% 75.3% 73.7% 112.0% 61.1%
Calcium 0.5 g 18.0% 92.7% 42.5% 44.8% 94.3% 115.0%

Sometimes 400g maize grain was given instead of 350g maizemeal.
Sometimes a portion of this has been soy-fortified corn meal.
Sometimes lentils.
Onions were used for the calculations, although other vegetables were also distributed, such as
cauliflower and radish.
150g January to March '93; 120g March to June 93; lOOg from July '93 onwards.
Except 1988-1992, when Faffa was distributed irregularly and often as a substitute for grain.
From WHO (1997) 'Nutritional requirements in emergencies', extract from The
Management of Nutrition in Major Emergencies, WHO, Geneva (NUT/96.16 Rev.3).
These calculations were made using the "NUTCALC software (Epicentre & Action Contre La Faim,
From a diet whose iron is of low bioavailability.

2 Case-study: Bhutanese refugees in Jhapa and
Morang Districts, Nepal (January/February 1997)

1 The context few days for most in lorries and buses. The Terai
is the strip of land which adjoins India along
Nepal's southern border. It is part of the Gangetic
1.1 Bhutanese refugees in Nepal plain, only a few hundred feet above sea level.
Bhutan is a small landlocked country bordered Until the 1950s and 1960s the Terai was jungle
by China in the north, and by India to the south, and sparsely inhabited, as it was a highly
east and west. It existed in isolation until 1952, malarious zone. The Jhapa district was cleared
when domestic reforms were initiated and ofjungle and peopled almost entirely by recent
relations with other countries established. It is migrants from the hills and plains people from
believed that from the mid-nineteenth century the south. Jhapa and Morang are now the most
onwards there was a large influx of relatively densely populated districts (respectively 370
poor and marginalised people of Nepalese and 364 persons/ in Eastern Nepal, and
origin into southern Bhutan, which was heavily Jhapa has the highest proportion of area under
forested and infested with malaria. These cultivation (72.4%) (Shresthaand Ganai 1996).
people gradually cleared the forest and settled. The first refugees began arriving in south-
In the 1950s, when Bhutan implemented its east Nepal from southern Bhutan after October
First Five Year Plan, the Royal Government of 1990. Camps began to develop in 1991, but the
Bhutan (RGB) encouraged Nepalese immi- peak influx was during May 1992, during which
grants as extra labour for development month 11,000 new arrivals were registered. By
activities. The Nepali population in the south of July 1992 there were over 50,000 refugees regis-
Bhutan was granted full citizenship in 1958, and tered in the camps. Smaller influxes continued
many Nepalis were absorbed into the adminis- during 1993 through to 1995. In 1996 there
tration. In 1985, however, a new Citizenship Act were almost no new arrivals. Records of refugee
imposed stricter conditions which bore heavily numbers are maintained by the camp admini-
on Nepali Bhutanese; the government claimed stration run by His Majesty's Government
to have detected large numbers of illegal (HMG), in co-ordination with UNHCR. There
immigrants, and a strict census was undertaken. are several thousand refugees who live and work
Bhutan's sixth Five Year Plan (1987-92) in Nepal and are not registered as camp residents.
included a policy of 'one nation, one people' At the time of this study, there was a total of
and extended the traditional Drukpa dress code 91,801 registered refugees (UNHCR Kathmandu
and etiquette to all; contraventions were met 1997). They lived in seven camps in total, six in
with a fine or imprisonment. Then in 1989 the Jhapa District and one in Morang District. The
teaching of Nepali was discontinued in camps of Beldangi 1, Beldangi 2, Beldangi 2-
Bhutanese schools, adding to the southerners' extension, Goldhap, Khudunabari, and Timai
sense of cultural marginalisation. are in Jhapa, and Sanischare is in Morang.
Dissent grew among the southerners, the Timai and Sanischare were settled spontaneously
People's Forum for Human Rights (PFHR) was and so appear more overcrowded than the
formed in Nepal in 1989, dissidents were planned camps, the first of which was Goldhap.
arrested, and in 1990 the Bhutan People's Party The three Beldangi camps, all adjacent to each
was formed in India. In Bhutan there were mass other, constitute the largest single concentration
demonstrations, unprecedented in the kingdom's of refugees in Nepal, with a combined
history, and many arrests, and dieflowof refugees population ofjust under 45,000, which is by far
out of the country began, reaching a peak in the largest community in Jhapa district.
1992. They brought with them many allegations At the time of this study the UNHCR pro-
of torture, brutality and rape (Hutt 1994:12). gramme was in the 'care and maintenance'
The refugees crossed Sikkim in India and phase, the emergency phase having finished in
arrived in the south-eastern corner of the Terai late 1992. The international non-government
in Jhapa and Morang districts, a journey of a organisations (INGOs) have been recently

Acceptability and use of cereal-based foods in refugee camps

gradually reducing their direct assistance, and citizens in the Terai who depend on such labour
there has been an increased emphasis on refugee for subsistence. Socio-economic surveys conducted
participation and self-management of camp in recent years have produced equivocal find-
facilities. The RCU (Refugee Coordination Unit) of ings on the negative impact of refugee labour on
HMG was established in mid-1992, under the die local labour force (Thacker and Sharma 1996
Ministry of Home Affairs, for the purpose of and New Era 1993). There may not be a surplus
registering and documenting Bhutanese asylum- of labour in Jhapa, because of the boom in
seekers in Nepal. In 1993 HMG established a construction and other economic activities in the
permanent presence in the camps through the area since the refugee camps were established and
appointment of Camp Supervisors as the international agencies arrived. Some say that
administrative focal point in each camp. the refugees have largely replaced non-national
The refugees in the camps continue to express (Indian) labourers, not local Nepali labourers.
their wish to return to Bhutan, and frustration It has been noted that the refugees are willing
persists over the lack of a durable solution. to work for less pay, as they receive a full ration
Between 20 and 30 refugee associations and back at the camp. This has led to an HMG policy
other entities have been established on their of trying to limit numbers moving in and out of
own initiative. In 1996, a number of demonstra- die camps. In this sense they are officially 'closed
tions and 'peace marches' were organised to camps', but there are no physical barriers, and
push for their right to return to Bhutan. sanctions against transgressors do not seem to
Bilateral government talks (HMG and the RGB) be extensively nor consistently applied. The
have been held since 1993, the seventh and last opportunity to work, albeit on a casual basis, is
round being held in Kathmandu in April 1996. certainly a factor which directly influences food
Discussions focus on the 'citizenship' or 'cate- availability and the possibilities for diversity of
gorisation' of the camp population, with RGB foods among the refugee population.
insisting that only a proportion are genuine
refugees; others, it is claimed, are economic 1.2 The refugee camp population
migrants. There was no specific result in April Ethnic groups: The refugee population consists
1996, except an agreement diat discussions should of ethnic Nepali, most of whom formerly lived
be continued (UNHCR Kathmandu 1997). in Bhutan. According to a demographic survey
conducted in late 1995 (Bal Kumar and Prakash
Proximity of towns and markets: The Terai is well Dev Pant 1996), a large proportion belongs to
served with buses and cycle rickshaws. Access to the Brahmin caste (26.1%) followed by Chhetri
towns and markets is not a problem in terms of (18.1%), Rai (12.3%), Gurung (9%), Tamang
distance. (8.1%), Magar (7.6%), and Limbu (4%). This
Physical features of the camp and types of composition is not reflected in Nepal as a whole,
shelter: Available land is very limited. As of where the 1991 population census identified
March 1996, the total net settlement area of all the largest group overall as Chhetri (16.1%),
camps was 247 hectares (this included an area in followed by Brahmin (13.8%), then Magar,
Khudunabari allocated for any new influx). The Tamang, Rai, Gurung, and Limbu in
number of persons per hectare then ranged respectively smaller proportions.
from 190 in Khudunabari to 588 in Timai
(Shrestha and Ganai 1996). The refugee huts Religion: According to the same survey (Bal
and other communal facilities are constructed Kumar and Prakash Dev Pant 1996), 72.5% of
over a raised earthen plinth, with a bamboo the refugees are classified as Hindus and 20% as
skeleton covered by lattice walls of twigs or Buddhists; and in all the camps Hindus and
bamboo, and roofs of thatch or sandwich panel Buddhists form the largest and second-largest
(a thin layer of plastic between bamboo mats). groups. This compares with the Nepal 1991
Latrines are constructed similarly. One latrine census figure of 86.5% Hindu and 7.8%
is shared by two households, and they are built Buddhist. The minority religions are Christian,
in a row between the rows of huts. Animist, and 'others'.
Opportunities for working: With the proximity Language: Of the 1996 survey sample, 94%
of towns, and farms and rice paddies all around, reported that they spoke the Nepali language
there is ample seasonal opportunity for casual and 3.2% and 0.2% respectively reported
daily work. However, there is a numerically and themselves as Nepali-English and Dzonghka-
politically significant number of landless Nepali English speakers.

Bhutanese refugees in Nepal

Literacy: According to Bal Kumar and Dev Pant and about 45% under 15 years old. There
(1996), before arriving in Nepal, 60.6% of the seemed to be also a balanced proportion of both
Bhutanese were illiterate. Since arrival, the illiter- sexes (51:49 male:female). Of the total 1,529
acy rate has declined to 31.8%. The proportion households included in their sample, 174 were
receiving high-school education has risen from headed by females (11%). Most of their sampled
39% to 67%. The proportion of females attaining households had from 4 to 8 members; the range
classes 1-9 was 74.2%, compared with 25% while was from 1 to 17. The average household size is
in Bhutan (Bal Kumar and Dev Pant 1996). The usually quoted by agency staff as just over five
Oxfam non-formal education (NFE) programme members per household. It is not possible to
by 1996 had over 5,500 students participating, compare the demographic profile of the refugees
80% of whom were women. The CARITAS Nepal with that of the Nepali community in southern
programme of primary and lower secondary Bhutan before the exodus, without comparable
education had by the end of June 1996 32,416 data from that time; these may be available, but,
students enrolled in primary school, 695 refugee if so, there is no reference to them in secondary
teachers (student:teacher ratio of 47:1), and 696 sources reviewed by the researcher so far.
students studying at secondary level.
Leadership structure: There are 4-12 sectors in
Occupations: Although the vast majority of the each camp and 4-6 sub-sectors in each sector,
refugees are farmers, there is a sizeable number with an average of 80 houses in each sub-sector.
of educated and skilled Bhutanese, including Each sector has a sector head, and each sub-
doctors, teachers, engineers, civil servants, and sector a male and female sub-sector head. The
education specialists. A large proportion of the functions of this camp-management structure
camp-based staff of the implementing agencies are distribution, administration, and peace and
are refugee volunteers who receive incentives. The security. The structure resembles the admin-
vast majority of refugees claim to have previously istrative 'block' system which the refugees had in
owned land and to have been largely self- Bhutan. The block consisted of around 500
sufficient in food. According to the 1996 survey houses, headed by a Mandal. The Mandal was
by Bal Kumar and Dev Pant, almost 95% of the elected at open-air meetings; these elections
sample reported that they had owned land in were subject to manipulation. However, even
Bhutan when they left die country. However, a though the people often resented what some of
proportion of these, it emerged, had only rented, the Mandals did, the leaders seemed generally
not owned, the land. The survey estimated that to be well respected. The Bhutanese are used to
about 10% of households had been landless leaders having a strong influence over their
families. This issue of land ownership is complex daily life (Buggeland 1993).
and extremely sensitive, as it is linked to issues
of citizenship and rights of repatriation. Food practices: In South Asia there has com-
monly been a hot-cold classification of foods. It
The same survey looked at the proportion of
is primarily invoked in relation to sickness and
household heads involved at that time (early
restoration of balance within the body of 'hot'
1996) in gainful work (excluding household
and 'cold' forces, so that a febrile illness, for
work, but including the sale of poultry
example, would be treated by 'cold' foods such
products, milk or other household products) in
as rice or a milk product (Helman 1994:43). A
the camp. Only 3% were so involved, and an
specifically Hindu classification of food items is
'insignificant number outside the camp as well'
based on an assessment of which of three qualities
(Bal Kumar and Dev Pant 1996).
they possess: 'tamas', 'rajas', or 'sattva'. 'Tamas'
Age and gender composition: The age and is the quality which begets anger, envy, dullness
gender composition of the total population of or inertia, 'rajas' that which evokes sexual
camps is shown in Appendix 2(b), but this figure passion and general excitement, and 'sattva'
differs slightly from that quoted in section 2.1, that which promotes harmony and elevation of
because it was under verification at the time of the spirit. Hindu scholars are of the opinion that
the study. The figures indicate 50% of the total food should be 'sattvic', which means not
to be below 18 years of age, and around 13% of overspiced, and savoury, succulent, substantial
the total to be below 5 years of age. The 1996 and agreeable (Majupuria 1980-81:42).
survey findings (Bal Kumar and Dev Pant 1996) More specifically to the Bhutanese refugee
indicated that over all camps about 50% of the population, Unni Wikan's work on 'the girl
population was between 15 and 60 years old, child in Bhutan' gives some useful insights into

Acceptability and use of cereal-based foods in refugee camps

the social factors affecting nutrition. She makes 1993: A WFP/UNHCR food assessment mission
a number of observations from her field work in March included a report by the nutritionist in
among the Nepali Bhutanese households in the assessment team that thiamine deficiency
southern Bhutan. She notes that, while there existed (Gautam 1993). However it was not
was a variety of ethnic groups and tribes, there clear how this information was obtained, and
was a convergence in the way they perceived the observation was overlooked and not
and treated girl children. She considered that, followed up at the time (Robertson 1994).
rather than ethnic distinctions, it was factors like Cases of angular stomatitis and glossitis were
economic rank, caretaker's level of education, seen daily, but were not being recorded
and urban or rural residence which tended to individually on the surveillance form. An
affect life chances. Despite a preference for boys outbreak of suspected beriberi became
to be born, near-equal care was given to boys apparent in September, when an increasing
and girls in terms of nutrition, emotional care number of patients with neurological symptoms
and attention, health care and immunisation, were presenting at the health centres. An
and duration of breastfeeding (Wikan 1990) investigation followed, with refugees describing
A detailed study of household food allocation symptoms from May onwards. From October,
in Nepali families (non-refugee) in rural Nepal weekly surveillance figures were kept
in 1989 indicated preferential treatment accorded (Robertson 1994).
to adult men but not to boys. Baker's work in This outbreak occurred almost 18 months
Timai camp in 1994 indicated that there was after the peak influx of the refugees. In
inequality of food intake, in contrast to Wikan's December a SCF nutritionist investigated the
impressions. Baker lived in Timai for six weeks reported outbreak in relation to possible
and measured and weighed a single day's food predisposing factors, including the general
intake in 23 selected households of a total of 149 ration, food habits, and trading, as well as
individuals. Her main findings were that children factors of age, gender, profession or caste, the
were given a higher ratio of proteinxarbohydrate effects of pesticides, alcohol, and other food or
than adults, that males took precedence over drink interactions. Her study (Robertson 1994)
females at all ages, and that, apart from vitamin failed to identify any social factor such as caste or
C, overall the vitamin and mineral intake was gender which predisposed refugees to
well below the WHO-recommended daily developing signs of micronutrient deficiency.
amount (RDA) set in 1974 (Baker 1995). Numbers of cases varied significantly from
camp to camp (highest rates in Beldangi 2 and
1.3 History of malnutrition and extension), and a relation between cases and
micronutrient deficiencies time spent in the camp was identified, as follows
This section describes a pattern of acute mal- (Robertson 1994):
nutrition in the first few months following a
Time in camps (months) % of cases
major influx, and then an outbreak of beriberi 18
months after the influx. It summarises investiga- <6 5.4
tions around that time and interventions imple- 6-11 18.5
mented. Nutritional status at the time of the study 12-17 41.5
was satisfactory, but a small number of cases of 18-23 23.1
vitamin B deficiency were still being reported, 24-29 3.1
and this is discussed in the field study findings. 30+ 8.5
June 1992-December 1992: This period consti-
tutes the 'emergency phase' of the refugee It appears from sample surveys done in the
programme, this being the six months following camps during these months that there was no
the peak influx of refugees. During this time, concurrent high prevalence of protein-energy
protein-energy malnutrition (PEM), vitamin A malnutrition. There was no severe malnutrition
deficiency, and iron-deficiency anaemia were (<70% wt/ht) and very low rates of moderate
separately monitored, but not vitamin B and C malnutrition (c.1.5% <80% wt/ht) overall. This
deficiencies. The main nutritional concern during was in contrast to a higher prevalence of PEM in
these months was PEM; the surveillance data the previous year (Robertson 1994).
show a steady reduction in prevalence, so that During 1993 the general ration was set as
by the end of 1992 mortality and malnutrition currently (2.1.1), except for the following.
rates were at an 'acceptable' level (Marrack 1994). There was no blended food included in the

Bhutanese refugees in Nepal

general ration. The rice ration was 450gm from Beldangi 1 to test its acceptability (Upadhyay
January to September, then 430gm from October 1994). This was conducted and was reported on
onwards. Vegetables were 150gm from January two months later (UNHCR Jhapa 1994a).
to March, 120gm from March to June, and The acceptability pilot in Beldangi 1 lasted
lOOgm from July onwards. Robertson calculated for one month (February). The results indicated
that in fact the average amounts received good overall acceptance of Unilito among all
during the months from January to October age-groups, and no evidence of households
1993 were for rice, pulses, vegetable oil, sugar, selling or exchanging it, either within or outside
and salt, 401 gin, 57gm, 25.8gm, 19gm, and the camp. This pilot also coincided with the
4.5gm respectively; and for vegetables 125gm introduction of parboiled instead of polished
from January to March, 96gm from March to rice into the ration, and so the opportunity was
June, and 11 lgm from July onwards. taken towards the second half of the survey to
Fortified blended food, in the form of WSB look at the acceptability of parboiled rice also.
(wheat and soya mix) and Unilito (wheat, soya The main findings reported were as follows:
and maize mix produced locally in Biratnagar), • Observations in a total of 212 households
was being used, but only for supplementary revealed that >90% of households had complete-
feeding for malnourished children (<80%WFH) ly consumed their Unilito ration by the end
and pregnant and lactating women (2.3.1). of the week (the ration period) and there was
Robertson's study concluded that the refugee no evidence of its being sold or exchanged.
diet was monotonous and deficient in • There was no significant difference between
comparison with that enjoyed formerly by the age-groups in terms of the way that Unilito
population before leaving Bhutan. In addition was prepared and cooked. The study suggested
she identified a number of key contributory that the fuel issue rather than personal pref-
factors. These were as follows: erence influenced the cooking method (i.e.
• the exchange oidaal from the ration for the prohibition on firewood in the camps,
foods containing less thiamine (and other B 2.1.5). Also it was distributed to all members
vitamins), or for non-food items; of the household regardless of age, and all
• the loss of water-soluble vitamins due to said that all members of the household liked it.
washing rice and prolonged boiling of food • Two out of three of the respondents (almost
in water; all were women) were aware that consump-
• the gross restrictions on trade and tion of Unilito could prevent occurrence of
employment for the refugees, which severely beriberi and other micronutrient disorders;
affected their own food-acquisition they had received this information through
possibilities (Robertson 1994). the SCF voluntary health workers and
The recommendations of the study, in addition community health workers.
to increasing surveillance, raising awareness, • The proportion of respondents who washed
and providing health education, included theirricemore than once before cooking was 95%,
improvement of the variety and quality of the die majority of these washing it twice. Most (77%)
general ration, and suggested some options. said that they retained the rice water for use
These were providing the more acceptable in curry or daal, while the rest threw it away.
black and green daal (instead of the yellow), • Of 106 households asked, the majority (63%)
providing wheat flour or fortified flour in place claimed that, given the choice, they preferred
of a proportion of the rice ration, and providing polished rice to the parboiled rice. The main
parboiled instead of highly polished rice. While reasons given were that parboiled rice is not
it was recommended that blended food be given suitable for the sick, elderly, and small children,
to all pregnant and lactating women, as well as nor for performing religious rites. Some also
to all children aged under 3 years as a supple- mentioned its bad smell as a reason.
mentary ration, it was not recommended to be • The reporter observed that refugees were
included in the general ration, as the study prepared to accept blended food, as long as
findings had suggested that it would be 'unpopular it did not replace any of the rice ration.
as a general family food' (Robertson 1994).
Graeme Clugston of WHO visited in early July
1994: Robertson had suggested that the latter 1994 to review the situation. At the time of his
recommendation should be investigated visit, vitamin A deficiency appeared to be
further. The UNHCR Food and Nutrition controlled, with all children <15 years of age
Adviser proposed a pilot-scale programme in receiving vitamin A capsules every six months,

Acceptability and use of cereal-based foods in refugee camps

as well as women in the first month after delivery 2.1 Use and acceptability of cereal and
(one dose). Cases of goitre and anaemia were blended food
evident, while prevalence rates of PEM among
the < 5 year old children were low (Clugston 1994). 2.1.1 Food acquisition
Clugston concluded from available epidemi- The general ration: At the time of this study the
ological, dietary, clinical, and laboratory evidence basic ration provided by WFP was as follows:
and other considerations that the refugees had
Rice 430gm/person/day
experienced a severe outbreak of beriberi,
Pulses 60gm/person/day
pellagra, scui-vy, and other micronutrient deficien-
Vegetable oil 25gm/person/day
cies, both single and multiple. However, the
Sugar 20gm/person/day
intense training given to health staff in the last
Salt 7.5gm/person/day
few weeks of 1993 probably led to some over-
Blended food 40gm/person/day
reporting of suspected beriberi and other
deficiency disorders (Clugston 1994). The supplementary ration, also provided by
He observed that during the first six months WFP, was as follows:
of 1994, all reported cases of suspected beriberi 80gm/person/day
Blended food
appeared to have been neurological (dry 15gm/person/day
beriberi), although the reporting did not 1 Ogm/person/day
Vegetable oil
distinguish wet beriberi as distinct from the
neurological beriberi. The highest incidence In addition, SCF/UK provided for the procure-
rates of severe beriberi appeared to occur in ment and distribution of fresh food such as eggs
different camps from those with the highest for severely malnourished children, and the
incidence rates of mild beriberi, and no infantile Nepal Red Cross Society (NRCS) provided for
cases were reported. No explanation for the the procurement and distribution of fresh
differing rates was proposed. vegetables with a ration of lOOgm/person/day.
The decline in incidence during 1994, These include garlic, onions, and turmeric.
Clugston suggested, was due to the combined The general-ration food items are procured
effect of including parboiled rice and Unilito, by WFP. The rice and lentils are produced
green or yellow vegetables (rather than radish) locally in Nepal and/or nearby India. The
in the vegetable ration, active case finding and blended fortified food is either WSB from USA
management by health staff, and a vigorous (in-kind donation) or Unilito (wheat, maize,
nutrition education programme. soya blend) made locally at a flour mill near
Another report (draft) produced in January Biratnagar. The oil is imported refined
by UNHCR Jhapa offered another perspective vegetable oil, fortified with vitamin A. WFP
on the suspected outbreak of beriberi. It was a procures supplies through a number of local
study on the outflow of relief commodities from sources (about four at the time of this study),
the camps, and it speculatively linked the outflow who win a contract through sealed tenders
of dual with the outbreak (2.1.1) (UNHCRJhapa (competition is strong) and then obtain
1994b). However, there was no formal epidemi- parboiled rice from mills in Jhapa (and lentils
ological evidence of this causal association. from western Nepal and nearby India). The
supplier receives a schedule of requirements
1994-1997: A continuing (albeit low-level) from WFP, then contacts the mills, and sends a
incidence of vitamin B deficiency has been supervisor to oversee the process and loading.
reported during these years which has not been The loaded rice is taken directly to the camp
fully explained. Some aspects of this are warehouse. WHO's 2,000MT warehouse in
discussed in the field study findings. Birtamod is for imported items only. Empty
containers and sacks are brought to this
warehouse after distribution and subsequently
2 The field study auctioned (c.l sack = 12NR, 1 oil tin = 23NR).
The considerable sums of money thereby raised
The following information was obtained through are put into a fund administered by
interviews with key informants and from 30 WFP/NRCS, with HMG approval, for expenses
interviews with refugees at household level. of camp warehouses, distribution counters,
Details of both sources are given in Appendix 2(e). feeder roads to the camps, and projects of the
Note: blended food will sometimes be referred Refugee Areas Rehabilitation Programme
to below aspitho; see Glossary. (Nepal) (RARP).

Bhutanese refugees in Nepal

If an item is missing one week, or delayed, section. He verified the card, and then the
NRCS/WFP policy is to supply retrospectively recipient proceeded to the counters. As the
(except vegetables), but only up to one week recipient visited each counter, the amount was
maximum of backdated supply. There have been measured out with a purpose-made scoop
no significant delays in the last years, because measure, either a ladle or tin, which when full of
the pipeline is short and transport abundant. that commodity contains one individual 14-day
The general-ration distribution system is ration, so that the rice tin when full contains
implemented by NRCS, which has devolved 6.20kg, the lentil tin 840gm, and so on. This was
much of the responsibility to the refugees. In then poured on to a small balance at the counter
each camp there is a 200-300MT capacity WFP as a double check. For bulk commodities like the
warehouse administered by NRCS. The stock rice and lentils, there was a large balance either
turnover is very rapid. No more than a two- in the middle of the hut or just outside.
week stock of rice, for example, would be kept As the sub-sectors were called in turn by the
there. In each camp there is a paid NRCS sub-sector heads, there was little sign of queuing
storekeeper and a NRCS Camp Relief Co- or long periods of waiting outside, so that
ordinator to oversee the whole system. disruption of a family's daily routine was minimal.
The refugees in each sub-sector of the camp As all camps are very compact, the distance from
organise themselves into distribution committees. the distribution huts to their homes was never
This means that all adults in the sub-sector are great. People knew which day their next issue
eligible to be on the committee, and they take would be, as it is a regular 14-day cycle.
their turn according to a rota, so the committee There were two distribution huts. On the day
is not a limited, fixed membership. They are when the distribution was observed, in one hut
responsible for the actual distribution at the a sub-sector of 82 families (441 persons) was
different 'distribution counters'. The head of being served, a process which took from 9 a.m.
the sub-sector organises the 'committee' to to 2 p.m. Most of the distributors were men,
perform the tasks of distribution. These are though women are said to take part also, on a
transporting the requisite amount of food from ratio of about 30:70 women:men.
the camp store to the distribution huts nearby, No refugee during interview complained
arranging the sacks and tins at each commodity- about the distribution system, either sponta-
distribution counter, and measuring the neously or when directly asked. They complained
amounts for each recipient. The head of the about inadequate quantity, but not about the
sub-sector holds the folder containing each system. They said they knew how much they
household's ration card (with hut number, and were entitled to, because they knew from the
family members eligible), and checks the details scoops and they could read the check balance.
and amounts before the householders proceed Even those who were not literate said they could
to collect their ration. This means that all 'read' the kg balance. In addition, their sub-
householders are receiving their quotas from sector heads were present and checked what
someone they know well from their own sub- each family was entitled to. However, when
sector, and in time all get their turn at distrib- asked what quantity of rice each person was
uting similarly to people they know, all from the entitled to, they said 6kg (instead of 6kg 20gm)
same sub-sector. This allows for a great degree and insisted they had always had 6kg. In the
of transparency in the whole process. The course of subsequent feedback sessions with
members of the 'distribution committee' receive groups of refugee women, this was again stated.
no incentive from NRCS. A key informant said One group also said that it was not easy to
they want to be involved because they are bored, complain, because all involved were from the
and they get the respect of the community same sub-sector. The other group felt no need
through volunteering. to complain. For some it was felt to be difficult to
From observation of a distribution, it appeared check-weigh the amount, especially if no one
that the system worked well and everyone else was so doing, because this might look as if
seemed very familiar with it. It was taking place you did not trust the committee and the sub-
inside two or more distribution huts (depend- sector head. When asked, they said that the sub-
ing on the size of the camp), in which there was sector head at distribution was always the man,
a ration 'counter' for each ration commodity. and were not sure why this should be, unless it
The sub-sector head was inside with a file was something to do with literacy levels. In
containing all the family ration cards for his sub- discussion with relevant key informants from

Acceptability and use of cereal-based foods in refugee camps

the agencies involved, it was explained that speculated whether refugees have a high birth
people say 6kg because 20gm is such an rate as a way of acquiring more food through
insignificant amount, and the amount received the general ration. This is linked to the
may be a little under or over, depending on the commonly expressed belief that those
size of the grain in that particular batch. The households with a relatively high proportion of
scoop used by all had been designed to hold 6kg young children are better off in terms of the
20gm, and it was not feasible to weigh bulk ration food supply (2.1.1). In response to the
amounts to the nearest 20gm. question, the SCF representative explained that
Sometimes only one person would collect the the birth rate has dropped substantially from
ration, either the head of household, or the wife 47/1,000/annum in 1995 to 26/1,000/annum in
of the head of household if he were busy 1996, and that the previously higher rates were
elsewhere, or both of them. Sometimes they due to low awareness of family planning among
would get the children to help carry. There the population. He also compared the current
seemed to be no prohibition or stigma in any 30% contraceptive prevalence rate in the camps
family interviewed against the woman going to with a 2% contraceptive prevalence rate among
collect the ration. Refugee informants said that the population when in Bhutan.
of those collecting the ration, around 70% were
women. In one household interviewed there Food purchase: The source of cash to buy food
were two wives sharing the same ration card but was from sale of ration items (2.1.1) or from
working inside or outside the camp (2.1.5).
living separately in adjoining huts, and apparently
Examples of market prices at the time of the
there was mutual antagonism. These women
study are given in Appendix 2(c).
collected their food together, then divided it
between them according to which members of The refugees depend heavily on the general
ration for the provision of a staple cereal, but
the family on the ration card they were cooking
sometimes families buy rice either from nearby
for. In another, a household which had divided
markets or from fellow refugees in the camp. If
itself between two huts was still registered on
they buy inside or on the outskirts of the camp
one card, so they collected the ration together,
from fellow refugees, then they will tend to buy
then sub-divided it between them subsequently. bagara rice, as that is what is supplied in the
In one household visited, it was the adolescent general ration, and it is cheaper by 2NR/kg than
grandchildren, male and female, who were the bagara rice in the market, which is in turn
collecting the ration, as the head of household 3^4NR/kg cheaper than polished rice in the
was their 80-year-old grandmother. market. However, for special occasions such as
Another source of food is through the supple- puja, polished rice has to be obtained. Bagara
mentary ration issued from the SCF Feeding rice is not appropriate for this purpose. A small
Programme centre. This is blended food as amount of polished rice was supplied once in
given in the general ration, but with sugar, oil, the general ration in recent months for a major
and dried skimmed milk (DSM) added. There religious festival, but that was a 'one-off
are two types of supplementary dry ration. One distribution because WFP had some supply of
is for malnourished children (<80% WFH) and polished rice which had to be used. Other food
medical referrals; the daily ration is: items such as rice flour have to be procured for
puja in order to make selroti and various curries,
WSB/Unilito 80gm/ppd possibly halwa, and in enough quantity for the
Sugar 15gm/ppd
ritual requirements and to give prasad to all
Oil lOgm/ppd
invited guests.
When asked what food people bought as
The other pre-mix is given to pregnant and priority when money was available, almost all
lactating women. It is the same as the above, but mentioned vegetables first. The type of
does not include the DSM. The pre-mix is made vegetable depends to some extent on the
up in the centre, and then given to the season, and during the time of this study it was
recipients. The women recipients are told that saag, potato, cauliflower, and tomatoes to make
the pre-mix is for them and should be kept acliar. Small vegetable stalls operated in all the
separately; follow-up home visits are conducted camps, but the vegetables were cheaper in the
to monitor this, according to the nutrition staff. local market than in the camp. Some refugees
During the interagency feedback session at bought directly from villagers who bring
the end of field work, one of the agency staff vegetables inside the camp to sell.

Bhutanese refugees in Nepal

After vegetables, meat was often mentioned, it. In one house a large sack of rice was evident,
but it was a special purchase, most people saying which they said was their excess from the ration;
they would buy once in one or two months. At but they offered no explanation of how long it
around 50NR/kg for pork which would supply had taken to collect that amount. One family
one large family for one or two meals only, the interviewed at the time of their puja said they
same family said they would rather spend the had obtained the requisite polished rice by
money on vegetables which would last four or exchanging their ration hagara rice. Whatever the
five days. Milk products such as fresh milk, scale of rice sale and exchange, the aim was not
curd, and ghee were also often mentioned, to obtain polished rice for daily consumption.
obtained from nearby villagers, many of whom While a number of people admitted to
would have some cows and/or buffaloes. exchanging some of their daal ration, primarily
People did not tend to buy extradaal, because to obtain vegetables, none of them said that they
the daal, which is widely preferred to that in the exchanged the blended food, except in terms of
ration, halo dual, is too expensive, at around borrowing or lending a small amount to a friend
35NR/kg compared with 24NR/kg. Extra oil is or neighbour to help them out until the next
not purchased usually, but ghee is purchased distribution. This informal small-scale borrow-
and used for deepflyingto make halwa,puwa, and ing happened with all commodities. One woman
selroti. Tea (with sugar or salt) is taken in all said she would not exchange the blended food,
households at least once a day, so this has to be because then 'what would the children have for
bought. One small refugee shop was selling tea at breakfast?' Some key informants agreed and
80-1 OONR/kg. In the town market it can be less said that there was anyway no market for this
or considerably more, depending on the quality. product in the local community. However, an
Shopping did not seem to be an exclusively interview with a local family living near one of
male or female task, but rather depended on which the larger camps revealed that this family at
member of the family was available at the time. least purchased pitho fairly regularly. The wife
pointed to her current stock, which she had
Food exchange/sale: A study conducted in early bought for lONR/kg. She said she thought that
1994 resulted in estimated outflows of about few people were selling, and she had several
2.5%-3% of rice, 6%-8% of daal, and 3%-4% of relatives in the camp, so was perhaps in a better
oil. An estimated 60%-70% of this was going out position than others to obtain what was
to be exchanged for other food items, so actual available. Some of her neighbours said that a
sale was around 30$h40%. The report stated diat certain amount did go to the market, but it was
there was nofloodingof the market and minimal impossible to quantify. When this was discussed
effect on market prices (UNHCR Jhapa 1994b). with groups of refugee women, they seemed
Interestingly, this UNHCR survey (draft) report surprised and disapproving (in comparison
concluded that the main reason for the outflows with their reactions to discussion of sales and
was the 'standard practice of providing food on per exchange of rice and daal). They said that as far
capita basis'. This idea was raised during the as they knew it did not happen, so it must be
field study by agency key informants and some only very few who were so doing.
refugees, who suggested that those households
with a rice surplus were those with a relatively Food production: Although most refugees were
greater proportion of small children within. involved in agriculture in Bhutan and are
This study found that daal was still the item skilled in the cultivation of grain and vegetables,
most commonly exchanged by individual the number of people with a small vegetable
households. One reason given for exchanging garden adjacent to their hut has apparently
food was the monotony of having the same kind decreased in recent years. The main reason
of daal (chana daal for more than 18 months given was lack of space. Some said that the space
now). Daal was exchanged with villagers from allowed was not enough to produce anything
nearby, who brought vegetables into the camp. worthwhile in terms of meeting the family's
Rice was also exchanged on an individual level. needs, so it was not worth the effort. Others had
There was also said to be exchange on a larger tried initially, with seeds given by the agencies,
scale, in which a group of neighbours pooled but found that certain plants did not grow well.
their surplus, usually rice, and then traded it Also after one or two years, the teams who came
locally and shared the proceeds. No one in inter- to do fogging against flies and mosquitoes made
view volunteered this information, but die group them remove their garden fencing, to get access
of refugee women consulted said they had heard of for the fogging procedure; without fencing, it

Acceptability and use of cereal-based foods in refugee camps

was difficult to protect the plants from small the findings from the 1996 Bal Kumar and Dev
children playing around the huts. Pant survey discussed in section 1.2.
Another major disincentive cited was the It appeared that most people had access to
realisation that fenced gardens next to the huts some cash to buy at least tea and spices and extra
were a potential hazard, in that they made fire vegetables. It was not possible to quantify this
fighting difficult in the event of a house fire. At cash income, but it was usually not difficult to
first some of the seeds issued had been cucumber ascertain that at least one member of the house-
and beans, which are climbing plants; the public- hold was working for cash, sometimes outside
health workers said that these encouraged the camp. In fact most people, on admitting
mosquitoes, and only small bushy plants should this, stressed that the work was 'sometimes' and
be grown. Some families had therefore not every day. In households with a male head,
discontinued their vegetable gardens and built it was the men who worked outside the camp,
kitchen extensions instead in the space. A few but female heads of household were also working
were maintaining small gardens full of flowers. outside. It was male and female adults who
The relief agencies have given seeds over the worked outside, as the children and adolescents
years, the last time being in June/July 1996, were enrolled in the schools, and great value
when NRCS distributed vegetable seeds such as was put on education by the refugees.
okra, tomatoes, chillies and aubergines. One man said they had to go out to work
Preserved foods and pickles such as gundruk secretly, because they risked having their ration
and sinky were normally made and used in cut. They worked in thefields,depending on the
households in Bhutan, but here they cannot season, and in construction work in the towns or
make them, because of the cost of ingredients, breaking stones for the road work. For carpentry
the equipment needed, and the space to do the work, the daily rate would be higher than for
processing. labourers. The daily payment, which included
lunch and tea on the site, ranged from 30 to
Opportunities for paid work and income
50NR/day. The rate was also higher if the work
generation: The two districts of Nepal where
the refugees are living are part of the food- was some distance from the camp. One woman
producing fertile area of the south-eastern said that women got 5-1ONR less for the same
Terai. Some key informants reported that, as work. One woman interviewed, who was head
most local people work on the land, there is a of a household with children, said she worked
labour shortage for construction work in the carrying stones for road work or in the fields,
towns, and also some labour shortage in the but only when the children were not at school
fields at busy times of the cultivation season. and could prepare their own lunch. She also
Most refugees were agricultural workers in collected and carried her own firewood or
Bhutan, but there is also among them a sizeable bought it if she had no time. She was formerly a
number of teachers and other professionals, factory worker in Bhutan. The main employ-
some living in towns, but some remaining in the ment was field work in season, and construction
camps. Several of these have been able to find work in the nearby towns. All ethnic groups
work in the camps with the international appeared to be involved in such work, but some
agencies assisting the refugees; they receive castes such as the Untouchable Darji (tailor caste)
incentive payments currently ranging from 600 have specific jobs which only members of that
to 1200NR/month. caste will do.
The camp rules state that, if refugees wish to Several families interviewed were also involved
leave the camp for a maximum of seven days, in income-generation schemes of their own
they must obtain an out-pass from the camp devising. One man (Brahmin) had numerous
supervisor. If they leave without the pass and/or bundles of saag in the house. These he had
prolong their stay beyond the seven days, their bought in the local market about 7km away and
ration is suspended for one-two issues. It was was about to sell in the camp to friends and
not clear how these rules were enforced and the neighbours. He did not appear to have an
sanctions applied. Refugees were aware of the official stall. Some had small window flaps in
restrictions, but there appeared to be consider- their huts which served as a shop window for a
able traffic in and out of the camps. Agency staff few wares, such as, in one house visited, limes
often mentioned the numbers of refugees and ginger, but the woman interviewed denied
employed in construction work, particularly in she was selling the goods. Refugee traders
Damak. This observation does not tally well with within the camp are supposed to obtain a licence

Bhutanese refugees in Nepal

from the camp authorities. Another family distribution but available in the market) or tins
(Untouchables) had recently acquired a small with lids brought from Bhutan which used to
pig and were busy chopping up banana tree contain their ghee; in plastic bags; or in a sauce-
stem for its consumption. They planned to keep pan with a lid on. In one house the blended food
it for about three months and then kill it and sell was in the large aluminium water-container with
the meat to their neighbours and with the lid which had recently been distributed by the
money buy meat andghee. They said this was the agencies. One woman kept her plastic jar
first time they had tried it. According to camp containing blended food in a locked chest,
rules, no four-legged animal is allowed within partly to keep the dust and dirt off it and also to
the camp boundaries. On the outskirts of some keep it away from the children. While the rice
camps pigs are quite numerous, and it is not and lentils can be washed clean of dust, the
clear how many belong to locals and how many blended food cannot.
to the refugees. A number of informants said
that there is considerable collaboration between Measurement: The amount to be cooked was
local pig-owners and refugees in the rearing decided by the person who was to do the cook-
and management of the pigs. ing, almost always a woman. Most people used
either a small tin to measure uncooked portions
Some women were observed around the huts of rice and dual to be cooked, or the traditional
making straw mats, for household purposes for bronze measure, called a mana, which they say
their own houses or presumably also to sell to holds about 0.5kg of rice. Asked how they
neighbours. A few pandits (religious men) were decided how much to prepare, they tended to
rewarded in money or in kind for officiating at say that they knew what the family normally
religious ceremonies, and reading holy books at would consume, and balanced this with the
the small shrines which many people had in need to make die ration last for 14 days. Typically,
their huts. one woman repoorted measuring out four mana
Agency income-generation projects mostly for one meal in winter, this for a family of eight
involved women, and in some camps there were people; others said one 'mana' for two persons.
small-scale projects for people in need of social
rehabilitation following the traumas of Bhutan Another woman showed how she used a tin
and flight to Nepal. The large Oxfam project mug, three mugs for six people. She said that in
has been handed over to the Refugee Women's Bhutan she would have used four mugs, if they
Forum (RWF). It includes weaving, knitting of had had no breakfast, but usually there they
sweaters and baby blankets, all for sale within would have had some breakfast of sweet potato,
the camp only, and production of chalk (in one maize or wheat flour, and so then she would
camp only) to supply the refugee camp schools. measure a little less rice for lunch. Not all
The Oxfam project has been very productive, women agreed with this, saying that they would
but the market is limited to the camp, and so always measure out the same amounts in
supply now exceeds demand. Bhutan, whatever breakfast they had eaten,
because there they worked on the land during
2.1.2 Food preparation the day and always had good appetites.
Storage: When the food ration is brought to the The amount of dual prepared depended on
family hut, it is stored in a corner, sometimes in how thick the soup was to be. Similarly the
the kitchen or in the main sleeping area, in sacks amount ofpitho used was according to taste and
and old oil tins. Rice and dual are kept in sacks, also method of preparation.
inside tins usually. The problem of vermin was
not raised spontaneously by the people inter- Preparation: The rice has to be cleaned as the
viewed, perhaps because during the winter first stage in preparation. This involves picking
months there is less of a problem. In answer to out stones, husks, and other debris and is
questions about rats, people said that there were invariably done by the woman who is to cook.
some, but there was not much they could do The rice is spread out on ajanglo, a straw platter
about diem. The rats did not consume significant specifically for this purpose, and meticulously
amounts of the rice, but polluted it with their the cleaned rice is separated and the debris
droppings. One family mentioned that they collected to be discarded. The time taken
bought an effective rat 'medicine' from the town. depends on how much is to be cooked and how
The blended food is stored separately, for contaminated the rice is.
example in containers with lids, either screw- Before parboiled rice was introduced in
top plastic jars (not supplied by the non-food 1994, most refugees had not used it before.

Acceptability and use of cereal-based foods in refugee camps

Most of them had grown their own rice for During household visits, the kerosene stoves
consumption. At the time it was introduced, usually appeared to be in use, but almost all huts
there was a public information campaign also had a supply of firewood and a wood stove
stressing that it should be washed in cold water fashioned in the hearth and signs of a recent
only once, and the water saved to add to the daal wood fire. Some people insisted that the wood
or the vegetables. Some health staff in the camps fire was only to keep the children warm in the
cited vigorous washing of the rice as one reason mornings and evenings (it was the coldest time
for what they saw as a continuing (albeit low- of the year, January/ February). They said that
level) incidence of vitamin B deficiencies. the kerosene stove cooked faster than wood, was
All those interviewed said they washed the quicker to get going, and did not need constant
rice at least twice, sometimes thrice, mostly in tending as did wood, but most women said that
cold water but sometimes in warm water. They they cooked the daal on the wood fire because it
kept the water from the second or third washing took so long (up to two hours); if they cooked
if the rice had been fairly clean, but not if it had daal twice daily, the kerosene ration would not
been dirty and/or smelling. They tend to clean be enough. This applies even though they soak
and wash the rice as they have always done, the daal in water before cooking.
though one woman said she washed the rice A few people had a kerosene pressure stove,
more in Bhutan, as it was dirtier. Most women but these were not common. Some people,
said they kept the rice water for daal or curry, again the minority, had pressure cookers,
but some did not, remarking that in Bhutan mostly brought with them from Bhutan. One
they would have given it to the animals. family interviewed had bought theirs in Damak
Daal, like rice, first has to be cleaned: the for 550NR. In one house there was a pressure
stones and dirt are removed, and it is then cooker, but the woman was cooking her daal in a
soaked in cold water for a period of 1-2 hours. small pan on the wood stove. She said this was
The blended food requires little to no prepara- because she had been out to fetch water and so
tion and cooking is very quick (next section). had left the children in the hut, and she did not
Cooking: Cooking takes place inside the hut. like to leave them alone with a pressure cooker,
The 'hearth', as the cooking area is termed, is for safety reasons.
either in one of the rooms or sited in an Almost invariably it was the woman who
extension to the original hut, but under cover measured the food, and prepared, cooked, and
and accessible from within the hut. Almost all distributed it to those eating at her hearth.
hearths observed during the interviews were However, there were exceptions to this. In one
noticeably neat and clean. It was rare to see any household (Brahmin), the wife of the head of
sign of food dropped and left on the floor, or household and mother of several children was
dirty dishes. Most huts have a raised cement attending the Oxfam non-formal education
washing area with a drainage hole just outside session at the time of the interview. Her
the door, so waste water is flushed away husband had just cooked lunch and fed the
immediately down a pipe several metres long. children and was awaiting the return of his wife
Kerosene stoves have been distributed to all. to eat with her. In a household with an 80-year-
People were generally not familiar with these old female head, the adolescent male
stoves before coming to the camps, and some grandchildren, she said, often prepared the
people complained that they broke within a meal if she was feeling too frail. It was also
couple of months. Large families receive two. explained that men knew how to cook, because
There is a stove repair/replacement facility when they used to be far from home during the
managed by NRCS, who have field-tested new day working in the fields, they used to take raw
and repaired stoves and found no difference in food and cook lunch where they worked.
fuel consumption and cooking times. New wicks The main meals of the day are lunch and
are supplied every 1-2 months. Kerosene is dinner. Both meals included rice and daal
distributed from a pump station in each camp. and/or vegetable curry, depending on how
The ration is for a household with up to three much the household chose to make and what
members, 1 litre per person per week, then 0.5 was available. The daal is made into a soup,
litres per week for every additional member so which is mixed on the plate with the rice, lidaal
that, for example, a four-member household is not cooked, the vegetables are fried and then
would get 3.5 litres, a seven-member household water is added to make a soup to mix on the
5 litres. plate with the rice. If daal is cooked, and also

Bhutanesc refugees in Nepal

vegetables, then the vegetables are usually fried The other most common form of preparing
only, with no water added. It is rare that dual is the blended food was to mix it with a small
cooked and not vegetables, which are most amount of water, fry it in oil, and mix in sugar.
usually eaten at both meals; women said that if This resembles the way puwa is cooked and it
there were not vegetables at the meal, then the produces a similar consistency. It is then eaten
children would not eat it. with a spoon or fingers.
Apart from cooking the dual, the rice and The staff of one of the SCF feeding centres
curry cooking-time should be no more than 20- which distribute a blended food pre-mix (with
30 minutes. This partly depends on whether the sugar, oil and DSM) advise recipients to add it to
cook is using one or two stoves (wood and kerosene) boiled water and then cook for 5-10 minutes,
at the same time. In the demonstration cooking whether the blended food is Unilito or WSB.
observed for this study, dual was not prepared. Staff from another centre said they told
The rice was cooked and then taken off the heat recipients to just mix with water and eat. If they
and left covered with a lid, while the same fire advise cooking at all, it is for the Unilito, not the
was used to prepare the vegetable curry, which WSB. All discourage frying, one in-charge
took around 20 minutes. While the rice was saying that frying burns the DSM and destroys
cooking, the potatoes, onion, garlic and cabbage the protein. In their guidelines he could find no
were washed and sliced. Oil was then put in the reference to this or to the mode of cooking. The
karai, and sliced onions cooked to brown Bhutanese nutritionist interviewed recalled
quickly. Just before the sliced vegetables were that, at the time of introduction of the pitho into
added, the haldi was sprinkled on and mixed the general ration, samples were taken to Dahran
with the browned onions. The prepared Food Science College and to the Food Research
vegetables were then added, along with the Laboratory in Kathmandu, to try to ascertain
garlic, salt and chillies, which had been crushed losses through frying, but the analysis required
on the silotho, and it was stirred from time to was too complex. He said that people found
time for 5-10 minutes. The rice water was then they could tolerate the WSB by just adding
added to make the curry soup; after 3 or 4 water, but suffered indigestion and flatulence if
minutes the cooking was over and the curry the Unilito was prepared in the same way.
removed from the heat. This is eaten sometimes Small children often ate the blended food dry
with achar, but tomatoes and other ingredients with no cooking; a little dry sugar was added and
not in the ration are required for this, so most mixed in. It was common to see children in mid-
people said they did not have it daily. afternoon wandering around with a cup or plate
After soaking, lentils are boiled for up to two of this mixture, and eating with their fingers.
hours. This is despite the fact that halfway through The blended food has been used as a weaning
the cooking the lentils are mashed to speed up food. Preparation is similar to that for other
the process and to make a smooth daal soup. family members, but the food is less likely to be
Cooking blended food is quick and easy. fried first. It is served with a spoon and is given
When it was first introduced, the information from the age of 5 months upwards. One woman
disseminated to all was to boil water first, then said that in Bhutan they would not normally
take it off the boil, add the blended food, and give satho until about 9 or 10 months, but this
stir. This was to avoid losing any of the vitamins was because there she could use cow's milk and
and minerals through any cooking process. ghee in which to cook and soften the rice.
Many people said this was how they prepared it,
but that they tended to add sugar. Instead of 2.1.3 Intra-homehold fooddistribution
adding blended food to boiled water and adding There appeared to be considerable uniformity
sugar, some added it to their tea, which is always between households in relation to meal times
sugared. The amount added and the final and snack times. The routine was similar to that
consistency were largely a matter of personal followed in Bhutan, and it fitted around the
taste and seemed to vary within families. However, daily work pattern. Most people said they had
a few people, young women and particularly food as follows:
pregnant women, have complained of indigestion 6-7 a.m. khaja (breakfast)
from the blended food, and one key informant 8-10 a.m. lunch: rice, daal, vegetables
said that he had heard that people suffered 1-3 p.m. khaja
from flatulence if they only added the blended 5-8 p.m. dinner: rice, daal, vegetables
food to water and did not cook it.

Acceptability and use of cereal-based foods in refugee camps

The variations in times depended on family The person, almost invariably a woman, who
circumstances, so that if, for example, breakfast has cooked is the one who distributes the food
was taken later, then lunch would be eaten later from the pans to the plates of those eating at her
and so on. Lunch tended to be earlier if they had hearth. She does not divide and give all the food
not eaten pitho at breakfast, and likewise dinner at once, but keeps most of it back to serve as
was earlier if they had taken only tea in the plates become empty, until the eater signifies
afternoon. Daal and vegetables were not always that he or she is satisfied and wants no more.
taken twice a day by everyone. Some who ate Food is not refused to anyone. At the time when
daal only once a day said it was to ease the all are satisfied, then she can eat. If there is not
monotony, while others referred to the long enough left to satisfy her, then one option is to
cooking time required for daal. One woman was cook again, but this involves more time and, of
cooking buffalo meat for lunch while being inter- course, more fuel consumption. Some said that
viewed, and was preparing daal but no vegetables if they know there is little to go round, they
to eat with it. Only one family interviewed said share it accordingly, telling the children that 'we
they had rice three times a day (the head of must share the little we have', so that in this case
household had been working for an agency and the woman would also receive a share. When
receiving incentives for two years). Several said asked how they tried to make sure that all had
that they had pilho twice a day as a snack, but enough, one or two old people particularly said,
others said they only had it once a day, because 'You have to control yourself, and implicit is the
there was not enough for two meals. comparison with Bhutan, where this kind of
Breakfast or khaja (a word which means a control had not been necessary. One woman
snack taken either early morning or later in the said, 'If we ate here like in Bhutan, then we
day) is first taken around 7 in the morning; this would finish up [all the ration] in a week'.
would be earlier in the warmer months. They If food remained at the end of the lunch,
eat only something light at this time. When most people said they would save the rice and
asked about khaja in Bhutan, people spoke of fry it up for a snack if the children were hungry
eating roast maize, or roast soya bean or roast later in the day. One woman said she kept evening
wheat flour (all termed satho) or sweet potato in leftovers for children to eat before school the
season, with tea. Their home-made satho in next morning, but mostly people did not keep
Bhutan would consist only of one cereal or them beyond 2-3 hours. Only the leftovers in
legume; it would not be a mixture like that the pan would be saved, not food left on the
issued in the ration. The food taken as khaja was plates. Leftovers, if eaten, are classed as a snack,
not classified by the refugees as food in the same not as another meal. Some said they never had
way that rice was classified as food. enough in the camp to have leftovers, but when
Now they were in the camp, almost all said they had them in Bhutan they would either fry
that they took pitho as khaja in the morning with them up for eating soon afterwards, or give
tea. Some wouldmix it with hot water and drink them to the animals. Where daal was eaten twice
their tea separately, some would mix it with the daily, it was always cooked twice, not once only
tea, and some would fry it first, mix with sugar and saved until evening, because that would be
and eat this, taking tea separately. Almost all 'like a leftover and the family would not eat it'.
adults said they had pitho early in the morning Another snack cooked by some, but appar-
and that the whole family would take it. Tea in ently not very well liked, was khole. This was
the camp is usually black, while in Bhutan they described in terms of a snack khaja food, but also
would have added milk. Almost all said that they as a way of eking out the rice supply. Some said
liked the pitho, but some added that adults did it used to be made in times of scarcity in Bhutan,
not eat as much as the children, so as to conserve but people tended to express dislike of the taste,
supplies. One reason it was made in the quick and others said that their family members
way, by just adding water, was so that the refused to eat it. It is particularly distasteful,
children could have it quickly before going to they said, when cooked from bagara.
school. If made as a drink, then less pitho was When asked if the whole family ate together,
used than was used in the more concentrated the most common answer was yes. The reason
puwa-type form. Several people said how much given was often that, if they ate 'turn-wise', they
the children liked thepitho and tended to ask for would not all have an equal chance to eat their
it at different times during the day. Their ration share. One man said that if they ate 'one by one',
would therefore not be enough for 14 days, there would not be enough. However, a number of
depending on how indulgent they chose to be. circumstances were cited where the family did

Bhutanese refugees in Nepal

not eat together, and it was almost always the Some even said that if they ate polished rice
lunch-time meal which was affected. A common now, they did not digest it so well. Others said
reason for eating separately was the children's that as refugees they had little choice, and had
attendance at school; this had been the case also 'had to like it'. The parboiled rice used to have a
in Bhutan. The woman would leave their portion bad smell, but now it was not smelling. One key
covered in the pan until they came home. In one informant suggested that it was not that the
house, the family had just finished lunch early, smell had changed much, it was just that the
because the husband was going out and others refugees had grown so accustomed to it that
had to collect the kerosene ration; but the two they did not notice it any more. It takes the same
children who go to school had not eaten with length of time to cook as polished rice.
them. The mother had put their food (o one One reason for acceptance of parboiled rice
side before the family began the meal. They might be that it is commonly considered as a
would eat the food cold when they returned. main reason why the beriberi problem, prevalent
However, where the wife or husband was out, in the camps 2-3 years ago, is now not seen. The
then the other tended to wait for them to parboiled rice and improved cleanliness around
return, so that husband and wife normally ate the home were often linked as reasons for the
together. One old woman head of household improved nutritional situation. Thericeseemed to
said, 'If I ate alone, then my tears would start'. be associated also with generally improved
Those interviewed who had received supple- health and conditions. They eat it all the time,
mentary pitho as pregnant or lactating women and buy polished rice only if someone is sick or
or as medical referrals said they kept it apart for puja. One woman interviewed had a young
from the household supply of blended food adolescent daughter who was being treated for
from the general ration and ate it as directed. TB. She had not procured any special food for
They said they did not share it with the rest of her during the time when she was feeling par-
the family or mix their supply in with that ticularly ill at the beginning of the treatment.
obtained from the general ration. This was because, if she had bought polished
Women breastfeed up to 4 years. There appear- rice, she would have had no money for
ed to be no cultural constraint on breastfeeding vegetables, and she did not make khole because
in public; several women were breastfeeding her daughter did not like it.
during the interviews and continued doing so The daal provided at present is chana dual, as
when neighbours, including men, came into the it has been for at least the last two years. Several
hut to listen or contribute during the course of people expressed a preference for kalo daal,
interviews. They start giving solids at around 5- which is more available and cheaper in Bhutan
6 months.
than in Nepal. In Bhutan it was the town people
2.1.4 Food preferences who ate cliana daal, which the government used
The basic items provided in the general ration to supply to employees, but in the hills people
are all food types which are familiar to the refugees grew kalo daal. One reason for preferring other
and would have been staples in Bhutan. Further- types of daal was simply to break the monotony.
more, the rice and daal are procured locally, For this reason, several people said they exchanged
either in Nepal or India, and one type of some of the daal, but they exchanged it mostly
blended food provided, called Unilito, is made for vegetables, not for kalo daal, because I kg of
in a factory near Biratnagar, a one-hour drive chana would be equivalent to only 0.5kg of kalo
from Damak. There was no complaint about the (according to market prices at the time, kalo was
ration, except that some said it was not enough, not as expensive as this — Appendix 2(c)).
or at least certainly not as much as they were The blended food provided is commonly
used to eating, especially in families consisting called pitho by the refugees, or lito, which is short
mostly of adults and teenagers, and that it was for 'Unilito', the product mixed and fortified
monotonous. locally near Biratnagar. It is also called .s7i//to and
Parboiled rice was introduced in the camps in pans pitho. These terms are generally inter-
1994. It was not familiar to the refugees, many changeable, whether the blended food being
of whom had grown their own rice in Bhutan; provided at the time is Unilito or imported WSB
those who did not would have purchased from USA. However, most people interviewed
polished rice. It has apparently taken very little said there were two types of pitho. One was not
time for the refugees to accept it, and in some ground so well and often smelled of maize,
interviews people expressed a preference for it. while the other was ground more finely and did

Acceptability and use of cereal-based foods in refugee camps

not smell of maize. Most expressed a preference Although people in Bhutan, especially those
for the latter, only one or two saying they liked in rural areas, would normally use mustard oil
both equally. The one they said was ground well for cooking curry, they receive refined vegetable
was the WSB, which they had been receiving oil (vitamin A fortified) in the general ration.
through most of 1996, and this was sometimes When asked their preference, some said they had
distinguished by the term paus pitho, meaning now become used to the refined oil, but gener-
'nutritious flour'. It was distributed in schools ally oil was not mentioned very much as an issue,
and health posts in Bhutan, and women already except in terms of the insufficiency of the amount.
familiar with it said it 'gives lots of energy and Ghee was obtained and used instead of oil for
protects children from diarrhoea'. certain special dishes such as halwa and selroti.
A key informant suggested that the stated There appeared to be no special foods which
preference for WSB may be grounded largely in traditionally would be cooked or obtained for
'psychological' reasons, as the refugees knew from pregnant women, and nothing was prohibited.
the packaging that WSB was a foreign product, If anything, women said that they tended to eat
so consequently they would place a higher value on less during pregnancy, because their appetite
it than the one they knew to be produced locally. was not so good.
During the study a new consignment of Other commodities not provided by the
Unilito had arrived and was being distributed ration, which the refugees most commonly
again. Blended food was not classed as a food in procured for themselves, were mainly tealeaves,
the same way as rice and dual. It was often not spices, and sometimes dairy products, and meat
mentioned spontaneously by interviewees in from time to time. Non-food items observed
general discussion about food, but, when and mentioned by people included tobacco,
questioned about it, they were enthusiastic and mosquito nets, clothes, and shoes for the
obviously valued it. It was considered as a snack children.
food, to be taken between meals with tea, and it
seemed from direct and indirect questioning 2.1.5 Local factors influencing food practices
that all members of the family ate it, including Fuel constraints: Refugees are not allowed to
the elderly. However, in interviews the cut wood for cooking or heating purposes, while
emphasis did seem to be on the children eating the local people may collect dead wood and cut
it, so that, while most adults would take it at least wood in designated areas. As the refugees are
once, mostly in the early morning with tea, the supplied with kerosene and kerosene stoves, the
children would ask for it at other times during rationale is that they should not require fire-
the day and would be given it. One or two wood also. Collection and cutting is considered
people said that pregnant women complained a violation of the Forestry Act. Of those refugees
that it gave them some indigestion (the blended in detention under Nepali law, the second
food for pregnant and lactating women is pre- largest group of offenders are those charged
mixed with oil and sugar). with firewood offences.
Pitho seemed then to be eaten and enjoyed The no-firewood policy is intended to protect
almost universally, and people tended to say the forested areas, and also to reduce the risk of
that it was a healthy food and good for them. It fire. Wood could be said to be more dangerous
was not something new to them, as in Bhutan than kerosene, because the flames are less
one of the snack foods they would eat at controlled and children can pick up pieces of
breakfast wassatho, which they made by roasting lighted wood and move around with them.
and then grinding maize. For this they had a Nevertheless, in almost all homes visited, fire-
hand-grinder and they would mix the ground wood was in evidence, in varying amounts. People
maize with ghee. However, they did not eat this had either bought it or collected it themselves
every day, because they had alternative khaja 'from the jungle'. Both men and women said
foods available. One family would eat puri, for that they carried firewood. One old woman said
example, because the head of household had that she collected it, rather than her adult son,
been a government employee and as such because the Forestry department were more
received an issue of wheat flour. This family said likely to arrest a young man than an old woman.
they would also have tea with milk, not black tea Where kerosene is also used for cooking, it is
as they had in the camp. The milk was from usually the dual which is cooked on firewood,
their own cows; only people living in towns because it takes longer. Kerosene is also used in
would use powdered milk. the small lamps used by people in the evening,

Bhutanese refugees in Nepal

including the school children for studying. One to share the water taps, so strict Brahmins said
woman said that her kerosene supply lasted they wiped the tap before taking their water. As
about 3-4 days, and that it would last the whole mentioned above, polished rice of good quality
week if she did not cook dual twice daily, boil such as aomusoli has to be obtained for the puja
water, and make tea. offerings. However, it is parboiled rice which is
eaten on a daily basis by all. Both parboiled rice
Religious and 'traditional' observances: RWF and blended food have been partly cooked
and other informants said that caste-related before distribution. This is against strict
restrictions had started to ease in Bhutan and Brahmin rules and the distinctions between raw
had continued to do so in the camps. The main and cooked food, but they said they had no
example given was that Untouchable groups alternative now that they were refugees. Those
were now permitted to enter most houses, and wishing to fast had to try to obtain fruits and
everyone was all 'mixed' together. Strict milk products to take instead of rice, and so they
observances were more likely in homes where had to buy or exchange ration items to do so.
there were elderly people, but the young were However, fasting is usually on one set day each
all mixing together at school and in the sectors. week, is a personal option, and for most people
Similarly the RWF said that many of the is not total abstinence.
traditional roles had changed. Women now left
The classifications of 'hot' and 'cold' food
the home and went to NFE classes or were
(1.2) were not mentioned by people during
working in the income-generation schemes.
interviews. According to a key informant it was
The literacy rate among adults and children is
mostly older women who now placed much
higher in the camps than it was among the
emphasis on such classification and consequent
population when they were in Bhutan (1.2). In
food avoidance. Not all foods are classified thus
the school all ethnic groups are taught in the
in any case. From the ration food basket, it is of
same classes, and the teachers include some who
little importance, because the chana dual is not
have studied abroad.
classified as 'hot', as the 'red dual' and black-
It was difficult in individual interviews to get grams are. There are few food avoidances
any sense of these cultural changes, except in between castes. Most notable perhaps are some
some cases people would indicate neighbours members of the Brahmin and Chhetri castes
who had come into the hut during the interview who avoid certain meat (especially pork).
to see what was happening, and imply that in
Bhutan that would not have happened. However, Understanding of health and nutrition issues:
in discussion with groups of refugee women, the One reason why the acceptability trial and the
consensus seemed to be that, while there was promotion of parboiled rice and blended food
much more mixing and joining together in was successful may have been that both took
groups outside the home, the same rules and place at a time when there were many cases of
restrictions that had applied between castes in mild beriberi and some severe cases among the
Bhutan still applied in the camps, particularly in refugee population, and the fear of the condition
relation to food preparation and consumption. was an incentive to comply with whatever
So, for example, a woman from an Untouchable measures were being recommended to prevent
caste may join the RWF and participate in all it. Many refugees during interviews and in the
activities with the Brahmin and Chhetri women, subsequent group sessions associated the reduc-
but she will still not be able to enter that tion in incidence of beriberi with the introduc-
woman's home, and the Brahmin woman would tion of parboiled rice and improved cleanliness
not eat anything prepared by her. Other around the homes. Health-education messages
practices continue also, such as the woman's about sanitation have been a constant feature of
demonstration of respect for her husband and the community-health activities in the sectors,
older children by eating food only after they so they would have been continuing during the
have finished eating; and the convention which intensive campaign about parboiled rice. The
dictates that during her menses the woman remarkable level of cleanliness inside and
cannot prepare and cook food for the husband between the huts is evidence of the success of the
and older children. Another female relative or sanitation campaign, so one could infer that
neighbour does the cooking during this time. compliance with the promotion of parboiled
The camp conditions present some problems rice may be part of that same responsiveness.
for those trying to observe social and religious In the last year, the camp health workers
practices, but they are not insuperable. All have have been undertaking training courses in

Acceptability and use of cereal-based foods in refugee camps

responding to psychiatric disorders. They have pellagra, suggesting that officially its aetiology is
been trained to consider the symptoms of vitamin deficiency. In a local health centre in
burning, tingling, and numbness of the lower the nearby large town, the in-charge said that he
limbs as possible signs of anxiety and mild usually classified angular stomatitis under
depression. They said they often referred such 'toothache and other mouth complaints', not
cases for anxiolytic treatment and sometimes under 'avitaminoses and other nutritional
anti-depressants, if vitamin B treatment did not deficiencies'. It is not clear how well understood
resolve the problem. One or two made reference are the conditions which arise from specific
to a condition of 'vitamin dependency'. The vitamin deficiency, nor is it easy to compare
health workers were aware of the possibility of incidences of, for example, angular stomatitis in
considering alcohol abuse in the differential the camps and in the local population, because
diagnosis, but did not consider it as a health they are classified differently. While the refugee
problem in the camps, compared with its referral system is a single channel, the local
prevalence in the remote rural areas of Nepal. population can attend any one of a range of health
One case of suspected chronic mild beriberi facilities, such as private doctors, pharmacies,
was included in the purposive sample. The and alternative healers. The camp surveillance
health worker had given a history of a young figures suggest a continuing incidence of
adult man who had complained over the last few angular stomatitis, but a negligible incidence of
months of burning and tingling which had not suspected mild beriberi. Refugees, when asked,
resolved, even after treatment with oral and said that the vitamin-deficiency problems were
intra-muscular vitamin B complex. When the no longer seen as before.
man's home was identified, he was not present
Aid infrastructure and refugee organisation:
and did not appear, but his father-in-law was
The international aid agencies' humanitarian
interviewed. When questioned, he said he had
relief programme in the camps was already well
no idea what was causing his son-in-law's illness,
developed by the time the parboiled rice and
only that he seemed to be better for a time after
blended food were introduced in 1994. The
the injections. He said the man had not eaten
physical infrastructure, logistical capacity, and
dual, for three years, because if he ate it, he could trained workers in all the sectors such as
not walk the next day. He said that everyone sanitation and health were well functioning and
else in the family was eating dual. co-ordinated. In tandem with this, the refugee
Some cases of angular stomatitis were seen community was internally cohesive, and a
during the study, and cases continue to be system of leadership and representation had
recorded at the Basic Health Units and Health been developed. What this meant was that
Centres. The cases observed during the study refugees could rely on certain basic services, so
interviews were mostly children; the parents that in terms of food they could plan and budget
said that it was a common problem in the winter on the basis of some security that food items
months and was due to the cold. One child was would arrive and be distributed at the expected
observed during interview to be particularly times. The public information system, through
thin and quiet and had signs of angular stomatitis. the schools, health clinics, refugee organisations,
His mother said he had always been a bit weak, and the sector and sub-sector heads meant that
but the sores were a problem of the cold information could be quickly disseminated
weather, although no one else in the family had through the community. Thus the campaign to
them. She had taken him to the Health Centre publicise and promote the rice and pilho and
and they had given him an injection one month ways to prepare, cook and use them was instru-
before. She said he had an appetite but ate only mental in their rapid uptake and acceptance.
small amounts. Interestingly, this child and the The refugee committees and the sector and
others had been regularly left alone with young sub-sector heads had been actively involved in
siblings for most of the day in Bhutan while both the promotion campaigns and received the
his mother and father went to do road work. respect and compliance from the majority of the
Some of the camp health workers interviewed population which was necessary for success.
gave the cold weather and poor personal hygiene There was an apparent level of social inte-
as the causes of angular stomatitis, but recorded gration within the camps which is consistent
the cases as angular stomatitis on the case tally with Wikan's (1990) observations of this
sheet, where the condition is listed in the same population in southern Bhutan just prior to
section as mild and severe beriberi, scurvy, and their exodus. She noted that, despite the

Bhutnnese refugees in Nepal

cultural diversity of the different groups, there The refugee nutritionist who helped to
were innumerable versions of marital connections promote blended food and parboiled food said
between them and generally 'close interaction' that a field test of acceptance would be
between them. In the camps they were now all necessary. One possible reason why it might be
mixed together in very close proximity and less successful than the earlier acceptability tests
shared a common grievance — that they were for rice and blended food was, he suggested,
forced to become refugees—and a common desire that now there was less of a problem of
to return home. While this close communal avitaminoses, and so people would not be so
living and enforced lack of activity had motivated by, for example, the fear of beriberi
inevitably affected the way they now lived, the — which had been one reason for acceptance
refugee conditions did not seem to have led to and uptake of the newly introduced food items
any major observable changes in their basic in the past. This was echoed in several
patterns of eating, of commensality, religious interviews, where people recalled how they had
observances, and intra-household food distribu- been suffering until parboiled rice was
tion. However, given a situation where education introduced, and because of that 'we are safe'.
and literacy for girls and women has become A group of refugee women from the NFE
dramatically more accessible in the lastfiveyears, programme were consulted also and they said
alongside the economic and social upheaval that, if it were explained to them and it was
which refugee status represents, there will be good, then there would be no problem.
cultural change underway which only a more The vitamin/mineral mixture could be easily
deeply anthropological study could reveal. distributed every two weeks as one more item in
the general food-distribution system. A
2.2 Opportunities for cereal fortification standard measure could be used, as for the
other commodities. Given the community
2.2.1 Household cereal fortification health infrastructure and the potential for
The cooking method allows for an opportunity public information and mobilisation by the
to add a fortification mix, just as turmeric is refugee leadership structures, and the
added to the vegetable curry. Turmeric is said influence of the RWF, the messages about its
to have no taste, but is essential to give the curry purpose and use could be well disseminated;
a certain colour and thus make it appetising. It follow-up could be done by the sector health
is added either just before the vegetables, after workers and/or the feeding-programme staff.
the onions have browned, or after the
vegetables have been cooking for a few minutes. 2.2.2 Cereal fortification in-country
Therefore the effects on the micronutrient The process of parboiling rice is described in
composition of the mix of frying at high Appendix 2(d).
temperatures would have to be assessed. The milling of cereals is common practice all
Although turmeric is not added to the dual, a over Nepal, from small hand mills in the homes
mix could be added at the end of the cooking to mills of larger and larger capacity. The five
process oidaal also. largest wheat-flour mills in Nepal have a
A hypothetical question about adding a capacity of around 120MT/day. The flour mill
pepper-type mixture with no taste to the meal owned by Mahalaxmi Nutritious Foods (MNF)
during cooking was asked during the course of near Biratnagar, where the Unilito plant is in a
several interviews. Invariably the response was separate building, has a daily capacity of 40MT
that, if it was vitamins and good for them, and if flour, which ranks it as a medium-sized mill.
the agencies explained how to use it and how What the Unilito plant has which the other
much to add, then they would do it. They mills do not is a roasting oven for the grain, and
sometimes added that the agencies had done mixing machines which add the vitamin/
the same with parboiled rice, which they now mineral mix to the ground flour. MNF had the
used and liked. The agencies through the roasting ovens already, as they are a large
health people and the sector heads had manufacturer of biscuits, but they imported the
explained how to cook it and why it was good for mixing machine from India (300.000NR) at
them, so this would be like that. When their expense and had a second machine made
questioned about possible sensitivities implied locally to the specifications of the first. This
by the addition of something to food in the project began in 1993. The firm received no
cooking process, those interviewed said there subsidy from WFP or from UNHCR, but a
would be no problem. consultant hired by UNHCR helped them to

Acceptability and use of cereal-based foods in refugee camps

decide on the appropriate type of equipment The vitamin/mineral mix comes from La
(Upadhyay, personal comm.). Roche, Switzerland, packed in aluminium foil.
The Unilito plant employs 50-60 people Calcium and iron obtained from India are also
when operating at full capacity, and has a added. MNF procure the maize, wheat and
maximum production capacity of 20-25MT/24- soya. Recently the composition was changed
hour day. WFP requirement for the present from wheat 50%, maize 35%, soya 25% to 40%,
refugee population is 600MT for a five-month 40% and 20% respectively.
period. The plant also produces Unilito for The latest batch ordered by WFP (February
WFP's Primary School Nutritious Food Project 1997) is to consist of wheat, maize and soyabean
in Nepal. There is another industrial group in (50%, 25%, and 25% by weight respectively).
western Nepal which has applied for WFP This reduces the maize proportion to less than
support to start producing blended food in that before. Quality control is contracted out to a
area. At present, the Unilito costs WFP more subsidiary of an international firm, which inspects
than the imported WSB. The MNF says it could the product and provides quality-assurance
reduce the price if it had guaranteed orders for certificates. Specifications of Unilito are given in
large amounts consistently. WFP says it cannot the appendices of the summary report.
do this, because it does not have an assured
budget for the programme beyond a certain
period of time. Also if WFP receives a donation Conclusion
of WSB, it must accept it, and so for that time
Unilito is not required by them. At the moment Blended food was widely consumed and valued
there is no competition between local suppliers, by this population. Cereal fortification as such
MNF being the only producer in Nepal so far. was not an issue, because fortification of whole
The Unilito plant comprises a roasting oven rice is not feasible. However, the introduction of
which is 110 feet long, through which pass the parboiled rice made more micronutrients
whole maize and wheat grain and the soya beans available in the daily diet. The acceptance of
at temperatures of 160, 180 and 200 degrees both innovations, particularly the parboiled
during a 10-12 minute roasting cycle. The rice, seems to have been greatly enhanced by
roasted grains and beans are then processed the public information campaigns around that
with a hammer mill, a pulveriser, and a stone time and subsequently. In addition, most of the
grinder. They are then sieved through a turbo refugees had been aware of the outbreak of
shifter and plant shifter. 250kg of the mixture is beriberi in 1993, and this had proved to be an
then put into the mixing machine with the incentive to comply. The food types in the
appropriate amount of vitamin/mineral mix and ration were all familiar to the refugees, and the
mixed for 25 minutes. It is then packed in bags, net main complaint was of monotony, not poor
weight 25kg. The outer bag is of a re-inforced quality or unfamiliarity.
plastic, the inner is a polythene bag. Shelf-life of
the product is 5-6 months in dry conditions.

Acceptability and use of cereal-based foods in refugee camps

Appendix 2(b): Bhutanese refugees:
age and gender composition (all camps)

UNHCR January, 1997

AGE 0 4 5 17 18+ NOTES TOTAL
BEL.1 949 941 1890 2896 2919 5815 4112 3876 7988 15693
BEL.2 982 931 1913 3559 3614 7173 5193 4983 10176 19262
BEL.XT 546 525 1071 1863 1788 3651 2722 2504 5226 9948
GOL 426 396 822 1605 3101 2207
1496 1998 4205 8128
KHU 650 735 1385 2086 4243 2996
2157 2789 5785 11413
SAN 1788 1820 3608 3004 5888 4208
2884 4006 8214 17710
TIM 550 467 1017 1630 1601
3231 2247 2106 4353 8601
TOTAL 5891 5815 11706 16643 16459 33102 23685 22262 45947 90755

N.B. These are provisional figures which might
change as data is revised following verification.



Appendix 2(c):
Market prices at the time of the study

The prices below are in Nepali rupees (NR) and were collected ad hoc during the course of the field study.
Note the range of price for different quality rice and lentils.

market (i) camp(ii) camp (iii) camp (iv) market(v) town info (vi)
Musoli rice kg 15 14
Aomusoli rice kg 12 13 12 12-15
Musoli kg 12.5 15* 10*
Basmati rice kg 22
Chana daal kg 24 24
Moong daal kg 40
Rah ar daal kg 42
Masoori daal kg 40 40 38
Kalo daal kg 35
Suji flour kg 20
Mutton kg 120
Chicken kg 100
Pork kg 50
Buffalo kg 35-40
Egg each 3
Sugar kg 28 24
Tealeaves kg 80-100 60-120

Cumin kg 140
Turmeric kg 48
Garlic kg 32
Ginger kg 10
Potatoes kg 8 6 8 5
Tomatoes kg 10 12 10 12 6 8
Onion kg 12 12 10
Cauliflower kg 6 6 10 8 8
Saag bundle
Radish kg 3

i) Sanischare market is on the way to Khudunabari camp from Birtamod, about 7km distant.
ii) This was a small refugee shop in one of the sectors run by a refugee. He said he brings the goods from
Budabare (a nearby market). Stocks, inter alia, spices, sweets and packets of suji flour. No rice or daal was
observed for sale.
iii) This small vegetable market was well within the confines of the camp. The vendor, a refugee, said he brings
the produce in a cycle rickshaw from just across the border in India. This may be why his prices were a bit
less than in other camp markets where the vegetables have been brought from local Nepali markets.
iv) Another small vegetable market within the confines of the camp.
v) This local market is just on the outskirts of Beldangi 2 and consists of a range of goods including food and
non-food items. It has existed only since therefugeecamps were established.
vi) This information was obtained from refugee interviewees at household level. The 12-15 price range quoted
for polished rice was according to whether bought on the outskirts of the camp(15) or at the local market
(12). The tomatoes arefromthe local market.

* 15/kg is more expensive than masoli polished, but masoli is more expensive than the type of rice which the refugees
are receiving. It would be cheaper to buy parboiled rice than polished if bought on the outskirts, or inside the camp.

** Wide variation according to quality.

Acceptability and use of cereal-based foods in refugee camps

Appendix 2(d): The process of parboiling rice1

Parboiling, as the name suggests, is 'partial milling there is a minimal proportion of broken
boiling' before milling. grains. Thus the yield is higher: for example in
• Paddy (rice with husk) is soaked in cold/hot a typical rice mill the yield of parboiled rice
water or steamed. could be easily 70kg per 100kg paddy,
• Soaked/steamed paddy is dried in the sun or compared with around 60kg per 100kg in the
with a mechanical dryer. case of raw (not parboiled) rice, which would
• Dried paddy is milled, usually in a two-step yield a higher proportion of broken grains.
process: Low loss of starch during cooking
Step 1: The outer husk is removed by the Only a little starch is lost in the cooking process,
process called shelling. From this, brown since it is gelatinised and dried before milling.
unmilled rice is obtained.
Step 2: Brown rice is milled to obtain
polished parboiled rice.
Disadvantages of parboiling rice
Note that in traditional milling the process of
shelling and milling is carried out in one step,
with a greater loss of vitamins. Smell
If parboiling is carried out using cold or luke-
warm water, the rice has to be soaked for 36
Advantages of parboiling rice hours. In this period, fermentation produces a
smell, which persists after milling. However, if
the parboiling process is done at a higher temp-
Retention of vitamins erature, or by using steam, then there is no smell.
In the process of soaking/steaming, the water-
soluble vitamins, which are mainly in the outer Longer cooking and harder grains
layer of the rice grain, penetrate inwards. When The grains are harder than those of polished
the rice is dried and the starch is gelatinised, the rice, even after cooking. For people used to
vitamins get trapped inside. Thus in the process softer rice, acceptability could be a problem.
of milling — in contrast with ordinary white rice
— the vitamin loss from parboiled rice is
minimal. Notes
Higher recovery 1 Information obtained in personal communi-
Since parboiled rice has undergone a process cation from Dr Janak Upadhyay, UNHCR
called gelatinisation, the grain becomes quite Geneva.
strong. In contrast with white rice, during

Appendix 2(e): List of key informants and interviewees

AMDA: Dr Mukti Nath Bhattarai MrSoshil Lai Kama, MPS,
SCF Health Centre, Timai
HMG Health Centre, Damak:
Ms Basanti Mahji, RWF Secretary, Goldhap
Mr Ram Briksh Chowdhary
Ms Coma Koirala,
CARITAS: Mr Loknath Pokhrel, Secondary RWF Asst.Secretary, Goldhap
Education Co-ordinator, Jhapa MrB.Thapa, MPS,
SCF Health Centre, Goldhap
NRCS: Mr Udaya Regmi, MrTek Natha Humagi, in-charge,
Programme Manager, Jhapa
SCF Feeding Centre, Goldhap
Oxfam GB Ms Memuka Nepal, RWF Co-ordinator,
Mr Shiva P. Aryal, based in Sanischare
Country Programme Manager, Kathmandu MsTila Khatiwada, RWFAsst. Secretary,
Mr L.P. Dahal, Programme Manager, Jhapa Sanischare
Mr Jagdish Shrestha, MPS,
SCF/UK SCF Health Centre, Sanischare
Ms Claudia McConnell, Ms Chandrama Subha, RWF Secretary,
Field Director, Kathmandu; Beldangi 1
Dr Ephrem E.Emru, Field Co-ordinator, Jhapa Mr B.P.Chaudhary, CHPM,
Dr Ashok Sharma, RARP Co-ordinator, Jhapa SCF Health Centre, Beldangi 1
UNHCR Ms Durga Gurung, RWFAsst. Secretary,
Dr Savitri Pahari, Beldangi 2
Health Co-ordinator, Kathmandu Mr Dik Shivakoti, CHPM, SCF Health Centre,
Mr Errki Heinonen, Beldangi 2/ext.
Representative, Kathmandu Ms Sita Giri, RWF Secretary,
Mr Robert Cooper, Head of Sub-office, Jhapa Beldangi extension.
Ms Corinna Miguel, Protection Officer, Jhapa Household interviewees
Mr Peter Janssen, Field Officer, Jhapa 30 refugee households + 2 local households
Mr Dinesh Shreshta, Technical Officer, Jhapa (hhh = head of household)
WFP: Place in household
Mr Quazi Haque, Wife of hhh 15
Country Director, Kathmandu Male hhh with wife 4
Mr Bijaya B. Amatya, Field Officer, Jhapa Male hhh 3
WHO: Mr William Piggott, Female hhh 3
Representative, Kathmandu Daughter-in-law of male hhh 3
Second wife of male hhh 2
Mahalaxmi Nutritious Foods Pvt Ltd.: Daughter of male hhh 1
Mr Santosh Sharma, Manager, Biratnagar Son of male hhh 1
Ethnic groups
Chhetri 7
Key informants — camp level Brahmin 5
Gurung 4
Mr Chandri Mani Moktan (nutritionist), Magar 4
Bhutanese refugee, Beldangi 1 Rai 3
Ms Garima Adhikari, RWF Secretary, Bhujel 3
Khudanabari Tamang 2
Mr Narayan Shrestha, MPS, Darji (untouchable) 2
SCF Health Centre, Khudanabari Gajmer (untouchable) 1
Ms Uma Adhikari, RWFAsst. Secretary, Timai Shrestha 1

3 Case-study: Somali refugees in eastern Ethiopia
(May/June 1997)

1 The context concluded that destitution was an important
impediment to resettlement in Ethiopia; and
Information in this section is taken mostly from pressure on land, as well as lack of assets and
secondary sources and some from key alternative job opportunities, threatened a
informants. See maps (Appendices 3(a), 3(b)). sustainable resettlement.
The returnees in KB had received a
1.1 Somali refugees in Ethiopia resettlement package but had not moved from
The refugees live in nine camps in Somali the camp, expecting a second resettlement
Region 5 of Ethiopia. They are from a number package to which they believed they were
of different clans. Most arrived between 1988 entitled,2 and also benefiting from the presence
and 1991, as summarised by Farah (1994a): of refugees, for example by obtaining bogus
'The escalation of the civil war in Northwest ration cards and claiming shares in rations from
Somalia/Somaliland in the summer of 1988 refugee kinsfolk. It is also noted that the
drove tens of thousands of refugees across the returnee camp residents had spent so long in
border into Ethiopia. Most of these belonged to camps in Somalia that they had become
the Issaaq clan. They found refuge in the Aware habituated to camp life: 'The existence of small
camps of Kam Abokor, Rabasso and Daroor and clusters of mainly Darod refugee groups in
in the camps of Hartisheik A and Harshin. The Kebribeyah, some of whom were settled there
latter was moved in 1989 to become Hartisheik after the official disbandment of the returnees,
B. The victory of the Somali National seems to have undermined the desired outflow
Movement in early 1991 did not bring of returnees from the camp. Many of the
immediate peace and stability to Northwest resident population and some powerful families
Somalia. Intertribal conflict drove Gadabursi living in Kebribeyah village obtained refugee
clan members across the border in February cards dispossessed of the weak refugee groups
1991, and two new camps, Teferi ber (Awbare) — the Hawiye, Ogaden, Harti, and Marehan
and Derwonaji, came into being. The SNM also clusters. Many of the local returnees still based
sent Ethiopian Somali refugees, most of them in the camps were ex-soldiers or in camps in
associated with the Darood clan, important Somalia; like the Ogadeni returnees, they do
sections of which had supported Siad Barre's not aspire to go back to the austere rural life
regime, out of the camps in Northwest Somalia prevailing in the areas controlled by the
back to Ethiopia. That led to the establishment corporate groups' (Farah 1994b).
of Kebribeyah camp, also in the spring of 1991.'
The boundary between Somalia and Ethiopia
The drought of 1989-91 also drove local is long and unguarded, and Ethiopia has not
pastoralists and agro-pastoralists to the camps. exercised restrictions on incomers. People
Conflict in Somaliland1 in November 1994 led living along the border are nomadic and
to a further influx of around 90,000 refugees pastoralist; they cross the border at will, so that
into the existing camps. the residents of the camps have continued to
Some of the camp residents are 'returnees'. move freely between the camps and North-West
These are mostly Ogadeni Somalis from Somalia. The areas to which the refugees fled
Ethiopia who fled as refugees to Somalia in 1979 are the clan areas of the refugees' clans. The
and then were forced to return in the late 1980s exception to this is the camp of Kebribeyah,
because of insecurity there. The vast majority where the original homelands of many of the
were re-absorbed immediately by the local Marehan and Hawiye residents are in the south
population, but a minority registered for of Somalia. A significant number are ex-soldiers
assistance in the camps. A survey in 1994 (Farah of the Siad Barre forces.
1994a) estimated the population of Kebribeyah
(KB) to be around 50% returnee, but only 5% in Proximity of towns and markets: see Appendix
Hartisheik A (HA). At that time, the survey 3(b). In the refugee areas around Jijiga a wide

Somali refugees in eastern Ethiopia

range of goods is on sale; many come from management of ARRA. The water is considered
abroad into Berbera and Somaliland. The area of good quality, and no chlorination or other
is a significant transnational trade route, with treatment is administered.
Hartisheikh being a major marketing centre, In principle, all camps have a medical doctor
and Daroor second only to Hartisheikh on the in charge of health and nutrition activities, but
Ethiopian side of the border. Food aid has not there have been many gaps in provision. All
been distributed in Jijiga zone since 1994. except KB have in-patient facilities. Most health
However, a proportion of returnees from workers are Ethiopian, apart from a few nurses
Somalia who are cultivating in rural areas in and health assistants who are Somali-trained.
Jijiga (between 0% and 25%, depending on the Health and nutrition reports are submitted
community) are receiving food the monthly to ARRA and UNHCR. There have
camps. In this sense, food rations in the camps been three reviews of the camps' health-
make up an important part of the surrounding information system (HIS) in the last nine years,
rural economy (SCF-UK Jijiga 1996). in collaboration with Centers for Disease
There are regular bus services to and from Control (CDC) in Atlanta, USA. The HIS is
Jijiga. Hartisheikh has a thriving 'bus station'. peculiar to the refugee camps, and is not that
Several buses each day move between the camps used by the Ethiopian Ministry of Health. It is
and Somaliland. There is much mobility and difficult to interpret the information from these
anecdotal evidence of commuting for business reports, for a number of reasons. The absence of
purposes. adequately qualified senior medical staff in the
camps calls into question diagnostic accuracy.
Physical features of the camp and types of The denominators used for calculating rates are
shelter: The camp residents live in aqals, huts the official overall population figures, but these
made traditionally of skins and mats, but here do not concur with the actual total number
made of mats, cardboard, cloth, and plastic suggested by the registration last year of all
sheeting, and attached to a hemispherical frame children below 5 years for the blanket feeding
of wooden boughs. Next to the sleeping and programme (BFP). Furthermore, the camp
eating hut, there is usually a smaller kitchen hut residents constitute a mobile and fluctuating
(alaxuaat). There may also be a small pen for population and the facilities are open to all
goats. The compound is in some way coiners, including non-camp residents, but
demarcated and at least partially fenced with these groups are not disaggregated in the
thorn bush or other border. reporting.
Humanitarian aid: The Administration for Food distribution is the responsibility of
Refugee and Returnee Affairs (ARRA), a semi- ARRA, and a group system is established (2.1.1).
autonomous agency within the newly- WFP supplies the food and has recently decided
established Authority for Security, Immigration that 'monitoring practices need to be improved
and Refugees, has overall responsibility within and reinforced in order to include also
the government for refugee matters. ARRA also qualitative aspects of the operation'(WFP 1997).
manages refugee camps and settlements in co- The plan is, in co-ordination with UNHCR, to
ordination with UNHCR and runs education, instigate different types of monitoring, such as
health, and nutrition programmes in the food-basket monitoring, market surveys, and
camps. The only international NGOs nowadays household monitoring, to assess food
directly involved with the Eastern camps
specifically are CARE and Handicap The 'cross-mandate' approach, 1992-1994:
International. The history of aid in these camps is long and
CARE is currently responsible for the extremely complex. In 1992, after the
tankering of water from bore holes several 'emergency phase', UNHCR and the Ethiopian
kilometres away to Kebribeyah and Hartisheikh government introduced a 'cross-mandate
camps and towns. At the time of the study, policy' for the camps. In theory this was to
54,000 litres/day was being delivered to KB encourage dispersal of the camp residents and
camp, and 33,000 litres/day to KB town. eventual closure, by looking at rehabilitation
Hartisheikh A camp was receiving 178,000 needs of the entire Ogaden region in which the
litres/day. The water is delivered to water tanks camps were situated. According to van Brabant
in the camps and distributed by hose and tap (1994) this policy foundered badly for a number
stands from there by the refugees under the of reasons, but generally because it was a 'policy

Acceptability and use of cereal-based foods in refugee camps

without a strategy and without a plan'. He the most northerly, is occupied by Somalis from
claims it was very poorly co-ordinated with Djibouti and is not included in the field study or
other agencies and led to the withdrawal of in the general comments below.) Counting and
cards held by returnees in the camps without registration of the camp populations has always
any rehabilitation work or plan for them having been dogged by controversy, and results
been operationalised. This meant that many disputed. There appear to be two main reasons.
returnees delayed leaving the camps. Firstly the Somalis in the region do not
Meanwhile the refugees, alongside returnees, categorise themselves by the same criteria as
continued to receive assistance, and nowhere those used by government and aid agencies, as
were these inconsistencies more visible than in described above. There exists between the
KB camp. camps and Somaliland 'a cultural and economic
Arguably the confusion of expectations of continuum'(Hogg 1992). Secondly, abuse of the
refugees and returnees which dates from these general ration system and ration-card system on
years remains as a legacy in the camps today, a huge scale by some of the beneficiaries over
and continues to influence relationships the years has led to inflated numbers when the
between the camp residents, government, UN, numbers are calculated according to the
and non-government agencies. number of ration cards being presented. As
early as mid-1989 this problem was apparent:
Repatriation: Voluntary repatriation has been 'Registration for food rations remains chaotic,
discussed over the years, but renewed fighting with over 400,000 cards issued but food
in Somaliland in 1994 delayed plans in that available for only 200,000 — so that families
year. The on-going peace process in Somaliland who do not have more than one card per person
has led to relative stability, with towns like are receiving only half their basic needs through
Boroma, Berbera, and Hargeisa again the official distribution systems' (Shoham et al
attracting private investors and people 1989).
returning to rebuild their homes and restart
Recent anecdotal reports (from non-
businesses. As a result, in January 1997 the
refugees) of individuals holding 20 ration cards
government of Ethiopia and UNHCR signed a
and transporting their ration by truck over to
Memorandum of Understanding for the
Somaliland were overheard during the study
implementation of a pilot repatriation project to
North-West Somalia of up to 10,000 refugees
A revalidation exercise, involving a census
from the Eastern camps in 1997. This project
and issuing of new ration cards, was carried out
was under way at the time of the field study.
in September 1994. Prior to this, WFP was
1.2 The refugee camp population providing food to 225,000 refugees (out of
600,000 registered ration-card holders). ARRA
The camp population includes refugees from ceased registering refugees after June/July
Somaliland, returnee Ethiopian Somalis who 1995. The refugees who arrived after this time
were previously refugees in Somalia after the and have therefore not been registered are
1979 Ogaden war, and some locally displaced estimated at 12,500. After revalidation in 1994,
Ethiopian Somalis. The categories of refugee, the number of beneficiaries went down to
returnee, and local have largely been construct- 184,000. However, by mid-1996, at the time of
ed by the international aid organisations, but the Food Assessment Mission (FAM), the
'this categorization does not fit the logic of the number of beneficiaries had reached a peak of
Somali families who have been driven back and 275,813. The FAM recommended another
forth across the border in search of refuge from revalidation exercise to be undertaken before
violence, but also in search of economic the end of 1996 (this has not been done to date),
opportunities or means for survival' (Farah and meanwhile based food-aid requirements
1994a). If the categorisation of refugee, for planning purposes on a 20% downwards
returnee, and locally displaced is only partly (if revision of numbers to 230,000, pending
at all) acknowledged by the camp residents, revalidation (WFP/UNHCR 1996). In fact, as
then one assumes that the residents' own the total food basket recommended by the FAM
categorisation of themselves is heavily has not since then been distributed to date, with
influenced by clan affiliation (see below). only cereal, oil and some blended food having
There are nine camps in all, which includes been available, the figure of 280,000 has been
two Hartisheikhs, A and B. (The Aisha camp, retained by WFP pro tern. The recent pilot

Somali refugees in eastern Ethiopia

projects of voluntary repatriation by UNHCR Occupations/opportunities for work: Pastoral
will have had also a limited invalidation effect, nomadism has been the most common form of
as those who register to receive the repatriation livelihood in the harsh environment in most of
package have to exchange a ration card for it. the Somali territories from which the refugees
originate. It is still the prevailing mode of
Clan/ethnic groups: Somalis constitute production for most Somalis.
relatively speaking a physically and culturally
Agriculture has become increasingly
homogeneous group. They are usually
important in Somalia, particularly in the
categorised as 'hamitic' or 'cushitic', 'hamitic'
southern, more fertile, areas, but agro-
being a term often used in opposition to the
pastoralism is common in northern areas. In the
'semitic' speakers of highland Ethiopia and
large towns of southern Somalia particularly,
Eritrea. The 'clan system' of the Somalis and the
urbanisation has developed. In the region
intricacies of clan affiliation and networks have
generally, including both sides of the
been explored and commented on in many
international border, agricultural production
reports in recent years. Much of the detail is not
has encroached on the heretofore exclusively
relevant for the purposes of this study, apart
pastoralist areas.
from references to leadership structures. The
fact that the refugees in each of the camps are all The refugees remaining in the camps now
in their own clan areas, except for those of are usually characterised by aid workers as the
Kebribeyah, has direct implications for survival poor urban dwellers who have no economic
strategies. resources to re-establish a livelihood in the
towns of Somaliland. In addition to these there
In summary, the Somali nation is divided are some groups who have fears about their own
into six clan families, Isaaq, Darod, Dir, Hawiye, security in their home places; many of these
Digil, and Rahanweyn. The former four supported (or are identified with those who
powerful clan families are primarily pastoral supported) President Siad Bane up to his
and widely dispersed; the latter are largely demise. After years of living in camps, many of
agricultural, concentrated in the riverine the households are now headed and supported
region of southern Somalia. Within the clan by women alone. In communities of urban
family, the next important social unit is the origin, men often abandon their families when
'clan', each one associated with a territory. Next they are not able to support them, or when the
is the 'primary lineage', and within this the basic children are older and shelter constraints mean
political and jural unit of the 'dia'-paying group, that there is no privacy for the couple. At this
numbering from a few hundred to a few point the man may leave and find another
thousand closely bound agnatic kinsmen, and woman to live with. For those of pastoralist
counting from four to six generations to a origin it is difficult, because formerly the
common ancestor. The 'dia' is further cemented daughters would be married and go to live with
by the obligation to pay and receive blood their husband's family from the age of 13
compensation and payments regarding minor onward, and boys of 9 years and over would
offences, thus producing a cohesive and stable spend time with their peers, looking after the
social group (Farah 1994c). animals (Ms Halimo 1997).
Religion: Virtually all the population is Sunni According to a repatriation survey carried
Muslim. out in early 1994 and focusing particularly on
women refugees' attitudes towards repatriation
Language: Somali is a hamitic language with
and their experience of surviving in the camps,
some borrowed Arabic words which was
the refugees' main means of livelihood was
transliterated into the Latin script only in 1972.
remittances from relatives living abroad,
Thus the Somali culture was, until recently,
assistance from their nomadic extended-family
essentially an oral one. During the course of the
members in the form of supplying them with
field study, more than one word was quoted for
animals for selling on the market or animal
a number of items, reflecting the existence of a
products, selling of firewood and water, and
number of regional dialects.
petty trading, mostly in quad, a leaf with mild
Literacy: The standard of schooling was good stimulant properties, in the camps and across
in Somalia, and education is a major concern of the border (Abdulkadir 1994).
the camp residents (WFP 1997). The literacy The SCF-UK household food economy
rate of women is lower than that of men. analysis (HFEA), conducted in late 1996,

Acceptability and use of cereal-based foods in refugee camps

includes much detail of KB and Rabasso, one of defined in terms of kinship, based on patrilineal
the Aware camps. Some findings are descent from a common ancestor.
summarised here to preface and complement It seems that these structures have been
the findings of this study. transposed into the camps, according to Farah:
The HFEA divided the population of KB into 'There is a tendency among the Somali to
wealth groups, with an estimated 50-60% 'poor', reproduce their clan structure in the
20-30% 'middle', and 5-15% 'rich'. It is stressed established camps. In the most general sense,
in the study that these groups would not existing camps occur in the territories
correspond with wealth groups in other camps, controlled by major landholding clans, each of
where the residents tend to be better off. Unlike whom dominates the resident population found
many of the other camps' residents, the people in the camp which is located in its sphere of
in KB arrived with virtually no assets and so had influence' (Farah 1994b).
little capital to start up businesses. Many women At camp level, the corporate groups which
said they would be more involved in trade and make up the clan are settled in sections which
other high-earning activities if they had the are geographically defined. 'Each corporate
capital to start. (While clan differences may group tend to act as an independent social unit
prevent men from engaging in trading activities whose members collectively attend to matters of
which require distant travel to other clan areas, common interest — particularly those relating
this is not true for women.) The poor engage in to the affairs of the camp. The composition of
activities which require little capital, such as the refugee committee in the camps, where all
collection and sale of firewood, making major groups are commonly represented, in
charcoal, working for locals, and petty trade. general, in proportion to their numerical
In contrast with KB, Rabasso residents are in strength, indicates this trend. In addition,
their own clan territory; they were helped by corporate groups have their own acknowledged
local relatives and some arrived with assets (cash strong leaders, whose opinion on current
or vehicles, for example), which they used to set matters such as repatriation is crucial' (Farah
themselves up in business (SCF-UK 1996). The 1994b).
profile of Rabasso is comparable to that of In the camps, the administrative structure of
Hartisheikh, as observed during this study. ARRA is of a different origin from the
Age and gender composition: There is no leadership structure which exists among the
reliable recent demographic profile, for reasons camp residents. The ARRA Eastern Region
outlined above. Perhaps one of the most Refugee Co-ordinator sits in Jijiga; his staff
usefully indicative figures is that for current includes the Zonal Health Co-ordinator for the
numbers of children under five years of age, refugees, and all the refugee camp co-
registered in the blanket feeding programme ordinators, who are Ethiopians, not refugees,
(BFP), which was around 21,500 in October and are recruited directly by ARRA. They all
1996 (Tsegaye 1996). It is not clear what speak Somali and have accommodation and an
proportion of the population the <5years office in their camp. In each camp there is a
children represent, but one assumes that, where UNHCR field assistant, also with an office and
families are divided between the camp and accommodation and radio communication
elsewhere, it is the women and small children (shared with ARRA). A Medical Supervisor
who are most likely to be the more consistent recruited by ARRA is stationed in each camp to
camp dwellers. Women are engaged in much of oversee the health and nutrition activities. In
the petty trading and the more weighty business each camp there is said to be a Refugee
and trading ventures. Here as in Somalia, they Committee (consisting of men) and a Women's
are traders in their own right, keeping their Committee. The latter is headed by a
own purse, enjoying considerable financial Chairwoman elected by the others, and on the
independence from husbands. executive committee there is a woman elected
from each of the sections. It was not clear how
Leadership structure: The social structure active and influential these committees are, nor
which accompanied the prevailing nomadic and how power is brokered between the ARRA
pastoralist way of life is defined as a administrators and leaders within the
'segmentary' or 'lineage' system. This system population of camp residents.
does not include a centralised, hierarchical
chain of authority or anything resembling the Food practices: Traditionally Somali commun-
State or a judiciary. Social relationships are ities are close-knit, and cohesion is based on

Somali refugees in eastern Ethiopia

formal kin/clan ties and traditional obligations rising, and vitamin deficiency, notably scurvy,
such as protection, compensation, and sharing was prevalent. The general ration supply was
of resources (Oxfam 1996). Of relevance is how erratic during the first few years of the camps'
much these ties have changed over the recent existence. After 1991 the situation stabilised and
years of intense conflict and mass migration. there were no more reported outbreaks of
Some groups may have retained cohesive trad- clinical scurvy. Malnutrition rates in the
itional bonds more than others, so the degree of different camps have fluctuated, but are also to
reciprocity in terms of maintaining livelihoods be interpreted with caution in later years, as
in gross adversity is probably variable. One of some rates are obtained from surveys, and some
the traditional reciprocal arrangements would from screening in the feeding programmes.
be between a group of rural women who would The food basket was adjusted according to
milk their animals each morning and pool the the recommendations of FAMs. No blended
supply. Then one of the group, on a rota basis, food was included in the general ration after
would travel to the nearest market, sell the milk July 1992, until its reintroduction was
and keep the profit. In urban situations this recommended in the 1996 FAM. This
hakbet scheme would involve contributions to a recommendation was implemented in part
central fund. This is reportedly replicated by from January 1997.
some of the poor in Kebribeyah, where a group
of women pool lkg of wheat grain and each 1988-1989: The period offlightof the refugees
tended to vary, according to their regional and
month a different woman in the group receives
occupational origin, from two days to 30 days.
the pooled grain and sells it to buy non-food
Those who came most quickly were more likely
items or set up a small business (SCF-UK 1996).
to be the richer, urban dwellers (Tsegaye,
Trading at all levels is a major feature of Somali
pers.comm.). The refugees arrived in
reasonably good nutritional condition, but
Attempts were made for the purposes of the there was a clear deterioration in nutritional
study to find any published material concerning status of children over the next eight months or
the traditional diets and recipes of Somali so, with rising levels of morbidity and mortality.
people in Ethiopia or Somalia, but such material In response, SCF began a community health
appeared to be virtually non-existent. The programme in 1989 which was handed over to
following is taken from a report on a nutrition the Zonal Health Unit in 1992. During 1989,
survey of five Ogaden zones and Jijiga a few there was a total of nine wet feeding centres in
months before the first refugee influx: 'The Hartisheikh A and B.
typical Somali diet is of a grain, milk and sweet The general ration in the early months of the
tea. The preferred grain is rice, but often this is refugee situation included Faffa as well as wheat
too expensive so sorghum, wheat or maize will flour and oil. Occasionally red beans were
be used and made into porridge or injera. The included, but these were not distributed
amount of grain consumed depends on the regularly, and nor was the Faffa. Faffa was given
availability of milk; in the wet season milk is in because at that time the Relief and
plentiful supply, and may form the entire diet; Rehabilitation Committee (RRC), then the
in the dry season the animals produce less milk government relief authority, and the Ethiopian
so small animals, usually goats or sheep will be Red Cross had some available; so when the
sold and the money used to purchase grain. grain pipeline was slow, partly because of
When substantial numbers of animals are lost, continuing war in Ethiopia, Faffa was used as a
as in the last drought, the demand for grain will substitute.
increase'(SCF-UK Dire Dawa 1988). Acute malnutrition rates were high. Survey
Breastfeeding is universal, but once the data show high rates particularly during the
lactating mother becomes pregnant again, she first six months of 1989, compared with before
ceases breastfeeding, as it is considered to be and after. 'The prevalence of acute PEM
harmful to the child. increased to a peak of 23.6% in March 1989, and
was still 22.9% in May. In addition, in March
1.3 History of malnutrition and 1989, the prevalence of clinical scurvy (defined
micronutrient deficiencies as bleeding gums and painful and/or swollen
The nutritional status of the refugees on arrival joints) was estimated at between 1 and 2%. At
was good. Within months nutritional status had that time no information on mortality was
declined to serious levels, mortality rates were available' (Toole and Bhatia 1991).

Acceptability and use of cereal-based foods in refugee camps

During 1989, nutrition-survey report forms 1991: By this time there is virtually no reference
included sections for reporting bleeding gums to vitamin C deficiency per se in reports. The
and sore joints, but these sections were not SCF-UK quarterly nutrition report in February
included after this year. The main commun- noted that the food basket was adequate for
icable disease of concern at this time, apart from protein requirements and below the require-
acute diarrhoeal disease, acute respiratory ment for kilocalories, but 'iron and vitamin C
infections, and TB, was an outbreak of hepatitis. has remained well below the recommended
In 1989, there were 5,185 cases and 96 deaths level'. The report concludes that 'the general
from hepatitis according to records (Toole and ration has again been deficient in micro-
Bhatia 1991). nutrients' and recommended that 'supplies of
Cases of scurvy were first identified in the oil, blended foods and beans/pulses as agreed
first few months of 1989. According to a key for general ration distribution by WFP/HCR,
informant (Tsegaye), it was mostly children must be maintained' (SCF-UKJijiga 1991).
below 12 years of age who were affected. It may In this year the number of refugees increased
be that adult cases were less likely to present at with the opening of Teferi ber, Derwonaji, and
health centres, while many children were being Kebribeyah from March onwards. For a time,
fed daily and screened in the wet feeding part of the food stocks allocated for Hartisheikh
programmes. It could also be that adult men were sent to the new camps, until the status of
were less susceptible because they regularly the camps was recognised and a food pipeline
chewed quad, and, as the most mobile sector of established. Some of the new arrivals did not
the population, they had access to a wider range register, but joined up with relatives. The
of food. According to Tsegaye, the women were condition of the new arrivals was poor, and the
very anaemic, but not scorbic. There seemed to malnutrition rate was high. The emergency
be a higher incidence in Hartisheikh A than in phase in the three new camps continued until
Hartisheikh B. mid-1992. From July 1992, Faffa was no longer
A team was organised and trained to screen included in the general ration.
the children for clinical signs. In the wet feeding
1992-1994: During these years, the 'emergency
centres, vitamin C powder was mixed with
phase' had moved into the 'care and
DSM, which was given twice daily.3 The area in
maintenance' phase, with a strong emphasis on
which the camps were situated is poor in sources
possibilities for repatriation, and also attempts
of vitamin C, and was until then pastoralist by UNHCR to develop a 'cross-mandate
grazing land. However, during this year approach' through which locally affected areas
business and trading centres were being would be assisted and camp residents who
established and vegetables were coming from might be returnees encouraged to disperse and
Harar 2-3 times weekly at prices which many settle outside the camps.
could afford (Tsegaye 1997).
From 1990 to 1994 the monthly general
1990: In this year an anaemia survey was ration issued was an amount enough for about
conducted, using a questionnaire and 20 rather than 30 days. Food supplies were
calculating the mean HB levels of adults and inadequate because of the gross inflation of
children <5 years. This suggests that by this refugee population figures. In September 1994,
time the main concern in terms of micro- a revalidation exercise in all camps reduced the
nutrient deficiency was iron-deficiency estimated figures (2.2) and the ration cards held
anaemia. It is not clear how much it was by beneficiaries were more accurately reflecting
considered that vitamin C deficiency was family size and the correct entitlement.
implicated in the prevalence of anaemia. The However, the regular distribution cycle which
quarterly activity report of SCF-UK (September was promised from this exercise was seriously
to December) included the note that among the disrupted in the Aware camps only weeks later
children in the feeding centres 'suspected by an influx of around 90,000 people from
scurvy has accounted for approximately 30 northern Somalia in October. Most of them
cases per month'(SCF-UK 1991:a). went to the Aware camps, which had until then
The general ration distribution continued to been comparatively small (around 10,000
be somewhat erratic, and throughout this year population since 1988). This influx was
the ration continued to provide only very low managed better than that of 1991 and did not
levels of both micronutrients. result in 'emergency' rates of malnutrition in
the affected camps, although scurvy rates of

Somali refugees in eastern Ethiopia

2.5% were reported among those in the feeding • the absence of beans/pulses from the general
centres, and many of the new arrivals lost ration for a long time;
weight after time in the camps. This could have • irregular oil distribution in the general ration
been due to disease resulting from the use of (in all camps there was no oil distribution in
dam water by people who had been used to three of the last six months' distribution);
clean water, and also the privations of the camp • new arrivals without ration cards in some
in terms of inadequate shelter material and camps;
blankets (Tsegaye, pers.comm.). One direct • increased incidence of diarrhoeal disease in
effect of the influx was that effectively the the first three months of the year;
repatriation initiatives were now 'on hold'. • water shortages;
1994-1996: During this time, SCF-UK handed • increasing cereal prices and poor purchasing
over nutrition activities in the camp to ARRA power of the refugees, owing to devaluation
(March 1995), which became the lead implem- of Somali and Somaliland currencies.
enting partner of UNHCR. The only inter- The study also found poor levels (23-48%) of
national NGOs left in the camps were CARE supplementary feeding programme coverage.
(water tankering) and Handicap International. Recommendations included the following
Camp residents continued to move in and out of (ARRA/UNHCR 1996):
the camps all the time, visiting relatives and
• improvement of the general ration to include
engaging in business and trading.
a regular supply of oil and pulses;
According to the UNHCR Health Sector • admission criteria for supplementary feeding
Annual Report of 1995, malnutrition programme for the next three months to be
prevalence rates in the camps were <10% widened to include all children <85% WFL;
(<80% WFH), except for Kebribeyah (12.9%), • the provision of an extra lkg of pulses in the
Aisha (13.2%), Derwonaji (11.2%), and Daror dry supplementary ration.
In fact it was decided to implement a blanket
May 1996: A Joint Nutrition Survey feeding programme for a period of six months,
(ARRA/UNHCR 1996) was conducted (16th- for all children <5years; this began in mid-July.
31st). A total sample of 4,535 children under 5
years (< 110 cm in length) was weighed and June 1996: The joint WFP/HCR FAM
measured, and percentage weight for length recommended a change in the composition of
calculated. The proportion <80% WFL (weight the ration to come into effect in January 1997.
for length) and <70% WFL and the mean WFL The FAM found that, as only wheat and oil had
(MWFL) were then calculated. The results been supplied, these commodities were being
indicated a generally poor situation in all sold to diversify the diet and to buy essential
camps, according to the national guidelines, non-food items. 'A great proportion of the food
with a range from 15.2% to 21.1% <80% WFL; distributed ... goes directly to the market'
the highest rates were in Kebribeyah and (WFP/UNHCR 1996a). It was decided to
Derwanaji. In all camps most of the redress this by adding sugar and Famix and,
malnourished children were <3 years of age. whenever possible, supplying sorghum instead
Possible reasons suggested in the survey report of wheat (sorghum was being supplied at this
were as follows: time for a few months). Famix was recom-
mended because of its digestibility for children,
• low levels of fat and protein in the food and its low market value (WFP 1997). The
received in the house; recommended contents of the new ration were:
• scarcity of milk during the dry season
(improved since April); cereals 400gm
• the risk of the children not getting their full oil 25gm
supplementary food allocation because of sugar 20gm
sharing between family members. salt 5gm
These results compared unfavourably with the Famix 30gm
malnutrition rates recorded in July 1995, This amounted to around 1,840 Kcals/ppd,
especially in the Aware camps, where rates were compared with the former 1,570 Kcals/ppd, but
almost double those of the year before. Possible was recommended alongside a reduction in
reasons given for this in the study were the numbers of beneficiaries (which did not
following: happen) pending the revalidation.

Acceptability and use of cereal-based foods in refugee camps

The FAM report was critical of the food- less than would have been expected from the
distribution system in the camps, identifying the total population figures. The conclusion drawn
fact that the family group system was open to from this was that 'the beneficiaries of this
abuse, because only the head of the group of programme are the refugees who are
twenty families needed to be identified; photo- completely dependent on the UNHCR food
graphic proof of identity was not required, and assistance, and have little or no alternative
cards were easily forged or bought. In addition, sources of food or income' (Tsegaye 1996).
there were no independent monitors present at
the time of distribution, UNHCR had not been 1997: The therapeutic, supplementary, and
allowed to be present regularly, there was an blanket-feeding programme rations have been
absence of distribution reports, and physical changed recently. From March, the rations for
checks or 'food basket monitoring' had never supplementary and for blanket feeding are the
been carried out. same: a pre-mix of 120gm Famix, lOgm oil, and
20gm sugar per person per day. This means a
July 1996: Blanket supplementary feeding of decrease in Famix but an increase of sugar for
all children <5years in all camps was begun. the beneficiaries of the BFP; but for the SFP
This had been recommended by the FAM. The beneficiaries it means a considerable reduction
ration was now set as follows:- in Famix and oil content. One reason for
SFP: standardising the BFP and SFP rations is to
1.7kg Famix pre-mix/week + 1.5kg pulses/week make it simpler administratively to calculate
pre-mix constituted as follows: food requirements. The BFP is generally
200gm Famix/day considered to have been a success, partly
30gm oil/day because of the tightly organised registration:
20gm sugar/day only those <5 years who were actually present-
ed for registration were included. The general
BFP: feeling among agency staff is that it is the most
1.1 kg Famix pre-mix/week needy who come to collect the ration weekly, as
pre-mix constituted as follows: those who are well off cannot be bothered to
150gm Famix/day wait and queue for it. In this way a self-selected
1 Ogm oil/day targeting is effected.
October 1996: The impact of the blanket No specific cooking instructions have been
feeding programme (BFP) was assessed by the given with the Famix. It is assumed that it is
Nutrition Programme Adviser to ARRA, three mostly made into 'a porridge'.
months after it was initiated. It was felt that There is no recent history of micronutrient
overall the situation had improved in most deficiencies, apart from the endemic iron
camps, but improvement had been slower than deficiency, iodine deficiency, and some vitamin
expected (Tsegaye 1996). The assessment was A deficiency (reported verbally but not evident
not a survey, but a screening of all those who in the recent monthly written reports). The
attended to collect their Famix from the BFP current reporting system includes as categories
(and who were weighed every two months). It only anaemia and malnutrition, and mal-
was noted that, while sharing among the family nutrition is defined as wasting according to
food which was given for children was common WFL. The last nutrition survey to be carried out
in all camps, it was more pronounced in the camps was in May 1996. Plans to conduct the next
with the highest rates of malnutrition (according to survey have been affected by the priority given
feeding-programme figures), namely Kebribeyah to the repatriation project under way at the time
and Derwonaji. This was due to poor economic of this study.
status and pressures on families.
Another issue raised was that the refugees
were not happy with the replacement of relief 2 The field study
wheat by sorghum in the general ration,
because of its poor quality, insufficient quantity,
and low market value, which had led to a 2.1 Use and acceptability of cereal and
substantial reduction in their purchasing blended food
power. In the following account of study findings,
This report also noted that the total number responses of both the interviewees from
of children <5years registered in the BFP was Hartisheikh A (HA) and Kebribeyah (KB) are

Somali refugees in eastern Ethiopia

included. Most of the time was spent in KB. men supported the use of the wheelbarrow,
Where responses differed noticeably, the source especially as they might be away working and
is placed in parentheses (i.e. HA or KB). doing other business. One woman said that
women were the ones to collect, because, if the
2.1.1 Food acquisition
men did it, they would waste the food in buying
General ration/SFP/BFP/repatriation package: luxuries; but another woman said that not all
During the first five months of 1997, KB camp men were like that. Where the amounts are
received wheat five times, each month, but at small, they may carry them themselves or split
different times of the month. No sugar was the load between two of them; but the use of the
received. Oil was distributed in January, March, wheelbarrow was very common.
April and May, but not in February. Not all of The previous distribution had been on 9
the camp residents are entitled to receive the April, constituting a 44-day interval between
general ration. There are returnees living there distributions. Most women interviewed gave the
who received a resettlement package some time impression that they knew how much they were
ago but have not moved on (1.2), and they are entitled to on the day, but seemed unable to
not entitled to the general ration. All quote their entitlement. Those asked
registration had been stopped in 1995, but in individually had received from 10kg to 12 kg
KB a registration of those previously excluded per person, according to them; but, when asked
had been carried out in 1996 and should have again during the feedback session in KB, they
picked up and included all those entitled but up said they thought it was 9kg. Certainly the
to then excluded. Nevertheless some women commodities they could expect differed some-
claimed not to have a ration card; but their times. They knew they should receive the ration
explanation for this was unclear. (or 'gift' as it was translated to the researcher),
During the study period in KB camp, the but they did not know on which day of the
May ration was received. It arrived on Saturday month it would come; as one woman said, 'until
24 May and distribution was completed by late we see the truck we can't know the time it will
morning on the 26th. WFP had informed the arrive'. Some women said that, when the 50kg
Camp Co-ordinator a few days before. The only sack of grain was measured out with their tin-
commodities for distribution were wheat grain cup measures of lkg, then there was only 40kg
and oil. According to the ARRA storekeeper, in the sack, but no scales were used to confirm this.
121.9MT wheat were received, and the official One group of women (KB), found dividing
ration was 12kg per person for the month (i.e. up their oil in one of the sections, were using a
400gm/ppd). The grain was off-loaded from the 0.5-litre cup. They said each person received
trucks into the compound of the ARRA Camp three of these (which is double the official
Co-ordinator. The distribution system is based ration). They had not received sugar in recent
on groups of 20 families, who are grouped years, but one woman thought they had
according to family size, so that in each group of received it twice about three years ago. They
20 the family size is the same. The sacks are then had received Famix in the general ration only
taken by the group leader, sub-divided on the twice, and that was in February and March
spot, and given to the families. Individual
1997; each time the amount was the same (3kg
kilograms are not weighed, but measured out
or 4kg each time among those interviewed for
with a large tin cup which all say holds the
the study). During the feedback session with
equivalent of lkg. As oil is more time-
agencies in Jijiga at the end of the field-work
consuming to divide up, it is usually distributed
period, agency staff replied to queries about
on a separate day.
When asked, women said that they had no public information on entitlements that they
problem with the group system of distribution. assumed that after ARRA had been informed by
Most, if not all, of the group leaders are men, them about any changes in the food basket, and
but it can be men or women within the group about delivery schedules, this information
who collect the ration. The sacks of grain are would be passed on to the refugees through the
taken away by the family, often on a Refugee Committee. The last formal session
wheelbarrow, rented out by people from the held had been a workshop in 1995 to
town, at a rent quoted variously as lkg of grain disseminate information about administrative
or 1-2EB, depending on the distance. When aspects of food distribution, but nothing similar
asked why they did not carry supplies had been done with regard to changes in the
themselves, they said they were too weak. The food basket or entitlements of different groups.

Acceptability and use of cereal-based foods in refugee camps

Eligibility to be included as a beneficiary of their food in Somalia. This included grain, but
the SFP and BFP does not depend on holding a particularly factory-milled white flour. Pasta
ration card. The criteria are nutritional for SFP and rice were mentioned, but for pastoralists
(<80% WFL and pregnant and lactating outside the town this was more likely to be an
women); for BFP all children < 110cm are exceptional purchase for a special occasion like
eligible. circumcision or a wedding.
The food component of the repatriation Most people interviewed said they bought
package, for those who register to be repat- wheat from the market to make enjera and
riated and submit their ration card, is as follows: shuuro and that they had to buy, because the
Wheat 150kg per person ration was insufficient. One woman, however,
Pulses (jejeb) 10kg per person insisted that she never bought, because the
Oil 5kg per person ration was sufficient. In such a case, it was not
clear if the interviewee was for some reason
During the study period, women in HA were unwilling to admit purchase or whether she did
complaining angrily about the poor quality of have enough for her household's daily require-
the wheat grain given as part of this package. ments. Almost all those buying wheat were also
Food purchase: For some typical market prices milling it at the town mill at around 17 cents per
of chief commodities at the time of this study, kilo (2.2.1).
see Appendix 3(e). Prices are quoted in At the time of the study, the prices of wheat
Ethiopian Birr (EB).4 and sorghum were approximately equal, so
In 1996 SCF-UK conducted a household there was no great saving to be made by buying
food economy analysis (HFEA) in Kebribeyah sorghum, as there would be at other times of the
and Rabasso camps. The HFEA divided the year. Those who bought sorghum mostly said
camp residents into 'poor', 'middle' and 'rich' they bought it to make garo and shuuro and to
groups. In KB the HFEA team estimated that add to wheat to make enjera, the wheat
50-60% were in the 'poor' group (< 100,000 constituting the main ingredient. Some said
Somali shillings per month). These people they never bought sorghum. A few bought
spent approximately two-thirds of their small maize to make shuuro.
income on high-calorie foods (maize and sugar) In Hartisheikh women were buying
and more than three-quarters of their income spaghetti at 6EB/kg and rice at 4EB/kg, when
on food in general. 'Purchases of nutrient-rich and if, as one woman phrased it, 'the old man
foods are minimal. The proportion of income brings home enough money'; but these
spent on food decreases as income rises, commodities were not mentioned in the KB
although the absolute amount of expenditure interviews. Meat was bought too, if they 'have
rises'(SCF-UK 1996). money', the cheapest cut being cow or camel at
From the household interviews and around 2EB, which would just be for flavouring
observations undertaken for this study, it was purposes, or 10EB to make a family meal.
impossible to gauge how much cash people had Almost all said that they purchased milk, or
available, but it was obvious that there were they were assumed to do so: containers were in
wide variations. At the lower end of the evidence. It was high on the list of priority
economic scale, women talked of collecting purchases when money was available. The study
firewood and selling and buying food for a day took place after the onset of rains, so the price
or so with that income (variously quoted from was low, at around 0.5EB for a small cup (0.2
2EB to 4EB per bundle). Several women litre), which they said could cost more than
referred to buying their essential items such as double that in the dry season. Mostly it was cows'
sugar and tea on credit (qarde) from shops in the milk that was being purchased. In poor
town and repaying their credit with the money households it was likely to be bought
obtained by selling their ration in bulk after the predominantly for consumption by small
general distribution (for 35EB for a 50kg sack of children: 'children have to have tea with milk'.
wheat). They said they bought little by little One woman whose 9-month-old baby was
throughout the month, only small amounts for 'growing weak' and looked undernourished
a few days. said she was buying fresh milk for the child
Before becoming camp residents, many of every day at the moment (as well as Famix
those interviewed, more so those from an urban sometimes). One young mother (HA) was
background, said that they used to buy much of buying tinned dried milk for her 7-month-old

Somali refugees in eastern Ethiopia

baby. When asked why she bought dried milk Both the "poor" and "middle" groups sell a
instead of fresh, she said that she could not keep portion of the ration' (SCF-UK 1996). It is
milk fresh in the house for more than a day, but important to note that this price differential is
did not have money available each day to buy it seasonal; at the time that this field study was
fresh. The 400gm tin had cost her 15EB. being conducted, the market prices of wheat,
Another commodity routinely purchased as sorghum, and maize were about equal
an essential food item is sugar. Heavily sugared (Appendix 3(e)). Nevertheless, the HFEA
tea is taken as part of the meal, for example, estimates that the poor in KB spend a higher
poured over the enjera at breakfast to make it proportion of their cash on food purchase to
more satisfying, or consumed in place of lunch if obtain calories and improve palatability than do
necessary. the poor in other camps, because those in KB
A number of women said that they bought need to make up for their higher levels of ration
Famix in the market; others said they did not. sales.
When one woman was asked why she might use This field study confirmed that most people
the little money she had to buy Famix and not seemed to be selling wheat, rather than directly
wheat or even meat, she replied, 'I don't know if exchanging it for other commodities. The
it is good nutrition or not, but I buy it because it proximity of markets makes this easy. Most
is cheap'. Others mentioned buying Famix to people admitted to selling a proportion of the
make shurbad for women who had newly wheat ration in order to buy other food, and
delivered. non-food items. Quite often women said that
One woman (KB), asked about priority they sold a whole 50kg sack for 35EB. A few
choices when shopping, said that it depended denied selling, saying that they consumed it all.
on how much money she had at the time. For No one admitted selling Famix, saying that
example, if you had only 50 cents, you might as the children needed it, and the amount they got
well buy tomatoes to go with enjera, because that was not worth selling. However, there was a
sum would not buy you enough sugar for any plentiful supply of Famix pre-mix in the
purpose. For one large thermos full of tea, she markets in both field-study sites. When asked
said she would use around 1EB worth of sugar where it was coming from, the vendors in KB
and 200ml of milk. said it was from HA, and some of the women in
One woman (KB) said she bought a tin of HA said it was taken from the feeding centre to
tomato paste only if there was a sick child or a be sold. A camp official in KB said that the
pregnant woman, both of whom tended to have women sold it after receiving it in the SFP/BFP.
difficulty digesting shuuro with oil. Other food As most of the Famix on sale is pre-mix (oil and
purchases reported and observed were jejeb, sugar already added), most of it is coming either
which were on the market then following their from SFP/BFP beneficiaries or from the camp
distribution in HA as part of the repatriation feeding centre where the pre-mix is made.
package; white bread in HA (same price as in Many people in Kebribeyah spoke of
Jijiga at 1EB for a long stick); and deep-fried borrowing from relatives when the food was
wheat-flour snacks from the camp market (HA). finished. Others mentioned that neighbours
Non-food items were many and varied, gave them food, especially where there were
including soap, cooking utensils, plates and young children in the family. If the food is
cups, paraffin, charcoal, and firewood. Men requested by a neighbour, it is given on a loan
particularly bought qaad, which was readily
basis; but if it is freely given, repayment is not
available in large quantities.
expected. A number of commentators on the
Food exchange/sale/reciprocal arrangements/ camps have noted the role of gift-giving in the
gifts: A major finding of the HFEA was that only Somali community. It is of course difficult to
in KB has the price of wheat consistently been quantify. One example of this was a family of
higher than the prices of locally produced maize siblings whose parents had died, and a teenage
and sorghum. On average over the one year brother had become mentally disturbed. The
from mid-1995 to mid-1996, refugees could sell teenage girl who was interviewed for the study
their wheat ration and purchase 34% more was effectively head of the household because of
maize. 'Therefore, in KB only, it makes her brother's incapacity. She said she sold her
economic (and nutritional) sense to see relief wheat to buy special food for the boy, and the
wheat in the market, as this is a key element in neighbours then helped out, not only with food
the coping mechanisms of the poorer refugees. but with tending him.

Acceptability and use of cereal-based foods in refugee camps

Several women talked of obtaining credit which the camps were and where many of them
(qarde) from one of the town stores for groceries. went collecting firewood for sale was very poor
Through the month they would buy tea, sugar, for fruits; only one was local (gob), and others,
rice and such commodities on credit, then repay such as hohob and uneho, were found only deep
with the money from the sale of wheat. A in the Ogaden. Sometimes they were brought to
number of women use this credit facility to buy the market for sale.
jejeb to make samosas for sale in the market each In many of the small compounds within the
day. A key informant at camp level said that in camp, there were small hen houses and
Kebribeyah there had been an income- enclosures for goats.
generation scheme in 1996, involving ten
refugees, who were each given a lump sum to 2.1.2 Food processing
start up a grocery and a butchery business. Storage: Normally when not in the refugee
However, the businesses had not materialised camps, pastoralists and agro-pastoralists store
by the time of the study. It was not clear how harvested food underground for up to two
start-up was progressing. A key informant said years. However, the food they have prepared
that liakbet schemes existed informally between for consumption, or purchased, is stored in the
women (2.4), but a group involved in the aqal. In the local village, one half of the aqal is a
cooking demonstration, when asked about raised platform for sleeping and storing food,
liakbet, denied any involvement. but in the camps, sleeping and storage are at
floor level. In most aqals visited in the camps,
Food production: All the camps are situated at there did not appear to be large amounts of
the edge of the Ogaden in areas where, with a food stored, apart from whole sacks piled up in
lack of groundwater sources and rocky soil, some. The aqals vary in size; many are in a state
vegetation is scarce. The absence of wildlife, of disrepair, and the wall and roof coverings
forests, and rivers limits the possibilities of consist of traditional matting, UNHCR plastic,
collecting foods in the wild. The rains had relief-grain sacks, pieces of cardboard, pieces of
begun by the time the study began, and the cloth and other sundry material.
surrounding areas looked green. Between Jijiga It seemed that most women were milling only
and the two camps there was considerable small amounts of whole grain at one time, no
agricultural production, the main crop being more than 5kg. One reason for this is that
maize. storage of flour in the«^a/s is problematic, since
In both camps, maize was growing as well as dust and debris can pollute flour easily. One
smaller amounts of sorghum, some beans, sweet woman kept her wheat flour in a tin covered
potatoes and tomatoes. Mostly women said the with a cloth; she said it could be stored like that
maize was primarily for the children, who for up to ten days. Wheat flour and Famix were
would roast and eat it as xasiid, or else the xasiid often kept in plastic bags, suspended from the
was sold. Later the older maize would be inside wall of the aqal away from the floor dust
ground in the mill and shuuro made from it. and rats; others store Famix (usually kept for
When asked why she was growing maize rather only a few days at the most) in a saucepan with a
than sorghum when in Somalia she had mostly lid. The plastic bags are multi-purpose, and
grown sorghum, one woman replied that it was small twists of pink and blue plastic are used to
easier to grow than sorghum, that the garden sell, buy, and carry home small amounts of milk,
was not large, and that the children could eat tomato paste, and oil. Small twists contain also
the new maize after roasting it (i.e. in that sense, spices of different kinds. The whole plastic bag
it requires less preparation than sorghum). is used to carry lkg or so of grain or sugar.
They expect two harvests per year of maize, but Unfortunately it is not bio-degradable, and
one of sorghum. One woman who was growing remnants of this plastic adorn thorn bushes for
beans said she had not grown them before in miles around and litter the fringes of the towns
Somalia, but had bought and ate them and camps.
regularly. Here that type of bean was not Five-litre oil tins with ropes threaded
available in the market. In HA one woman said through at the top are very common. They are
she had grown maize over the years, but had no used for collecting Famix and flour from the
crop now, as she was waiting to be repatriated. mill, and also storage in the house.
Although women, when asked, tended to Whole grain is stored inside the aqals in sacks,
know the names of the chief wild foods available and other commodities in cardboard boxes
in some locations, they said that the area in from ration distributions. A few traditional

Somali refugees in eastern Ethiopia

storage containers were also seen, including prepared 1.5kg, whether it was for enjera or
gourds for small amounts of grain, and, in one shuuro (the weight of flour being less than the
large aqal, the woven wooden baskets which are amount of whole grain before processing).
slung from the sides of the camel for Another, cooking for 8 people, used 1 kg of flour
transporting bulk items. plus a handful of Famix to make the enjera
The prepared enjera dough is usually stored dough. Others quoted similar proportions for
overnight in the aqal, not in the alawaat, which enjera, one woman saying she needed 0.5kg
would be too warm. It seemed to be mostly more flour to make shuuro.
mixed and kept in a small plastic bucket with a lid. It was interesting that, when the women were
Small amounts of fresh milk are carried in asked what proportion of Famix was added to
blue plastic bags, and stored in tin cups which what proportion of water, their answers ranged
most women appeared to have for drinking tea widely (from 1:2 to 1:10), despite the fact that
and milk and for measurement purposes, there the cooking method and end product were
being three sizes (see below). For storage for consistently similar. This lack of accuracy
more than a few hours, cow milk is boiled, while perhaps reflects the easy and casual cooking
camel and goat milk is heated but not boiled. method, where either water or Famix can be
Water is collected and stored mostly in 20- added incrementally without detriment to the
litre jerry cans. Distances to water points are not outcome (3.1.2).
a problem, and there was no sign of congestion or
Preparation: The most commonly prepared
long queues around water points in either site.
dishes included those which required milling of
Measurement: It proved difficult to work out the grain to flour and those which required only
how much was being prepared per person, for pounding to dehusk the grain.
three main reasons. Firstly most women were Dishes requiring milling to flour: Enjera (or
claiming to be eating only grain with oil for canjeero, kimis or laxoox) is made from finely
meals, and not adding meat and vegetables. milled white wheat flour, either machine-milled
This may or may not have been true; but or hand-milled. When hand-milled, the wheat is
obviously, if meat and vegetables are prepared cleaned, pounded with the vwoye, and then
alongside the staple dish, the latter portion will winnowed with the masaf. The enjera flour can
be less than if it is the sole item of the meal. be mixed with a smaller proportion of maize,
Secondly, a number of women claimed that the sorghum flour, or Famix (as in the case of the
size of household for which she was cooking was cooking demonstration for the study); but,
larger than that on the ration card, one reason according to the women interviewed, it cannot
presumably being that invalidation of cards had be made from maize or sorghum flour alone.
not been carried out since 1994 (1.2). Lastly, a Wheat is the essential major ingredient,
number of women in KB made reference to according to these women. However, in the
'visitors' staying with them temporarily from local village, and according to key informants,
Ogaden. (It was not clear what was the status or enjera can be made from sorghum alone (but it
purpose of these 'visitors' and what should be eaten straight away, not kept for any
contribution was being made by them to the length of time once cooked). The flour for enjera
household food supply.) Certainly measuring is mixed with water to make a smooth, liquid
adequate quantities is not a problem for women dough. To the dough is added a starter
used to cooking for their families, unless (dhananis) of already fermented enjera. This
unexpected visitors turn up (2.1.3). mixing takes only 10 minutes or so. The
It is the person who is to cook who decides mixture is then covered and left overnight in
how much to prepare. This is always a woman, the house, which is cooler than the kitchen;
usually the oldest woman or the eldest when fermented in the morning, it is ready to
daughter. Many people used tin cups to cook.
measure out grain and flour. There are four Shuuro can be made from wheat, sorghum or
sizes; the largest, a galaan, is used to measure maize. It can be machine-milled or hand-milled.
1 kg of grain, and just over one litre of fluid. The When hand-milled, the grain is cleaned with the
next sizes hold approximately 450ml, 300ml, masaf, rinsed with water, and then pounded
and 200ml. with amooye. It is then dried for a few minutes in
For enjera the amount measured out was the sun and then ground by hand with the stone
usually equivalent to 200gm per person. For 8- grinder (shiide). It is then winnowed, the coarser
10 people, one woman said she bought and wheat grain being used for shuuro; the separated

Acceptability and use of cereal-based foods in refugee camps

finer flour is the enjera flour, which is later cooking said that in the camp it would feed
added to make the mixture thicker (see below). about four people for breakfast, because in the
The shuuroflouris then ready to cook. camp they did not have lunch, whereas in the
In HA most women were taking their wheat town it would feed more than this, because they
for grinding to one of the commercial grinding had other foods to go with it and they took
mills in the camp, not to the free grinding mill. lunch. When it is about to be eaten, if there is no
It seemed that, while women in KB also used the sauce or meat to go with it, either oil or sweet tea
commercial mills, more of them were pounding is poured over it. If meat is available, it is fried
and grinding their own flour than were doing so and mixed with vegetables such as potato and
in HA. A number of factors seemed to influence red tomatoes to make a stew (sugo).
whether women used the mills or not (2.2.1). A less common dish made similarly is bur, a
Dishes requiring de-husking only: Garo can be type of bread made from dough which is not
made from wheat, sorghum, or maize. The fermented, but this was not observed during the
grain is not milled to a flour. To prepare garo, study, and women did not mention it
the grain is first winnowed with the masaf to spontaneously.
remove the chaff, then rinsed with water to Another favoured dish isshuuro. Theshuuro is
clean, If garo is made from sorghum or maize, gradually added to boiling water and stirred
the grain is then pounded to remove the 'eyes', continuously at the beginning of the process;
but this is not necessary for wheat, so then it is covered and left to boil. The cooking
preparation time with wheat is less. time depends on the amount. For the study
Karis can be made from wheat or sorghum, demonstration, only 0.25kg shuuro was added.
but not maize. The grain is not milled to a flour. After about 25 minutes, a relatively small
To prepare the grain for karis, it is first cleaned amount of enjera,flourwas added and stirred in
with the masaf, then rinsed in water. The wet to make the consistency thicker. The cooking
grain is then put into the mooye and pounded. time in all was about 40 minutes; the women
The moist pounded grain is then dried in the said that when cooking for the family it usually
sun for a few minutes and then winnowed to takes twice as long. Apart from the enjera,
remove the husks or buushe, which are then nothing was added during the cooking process,
given to the chickens. The grain is then added to but butterfat or oil is usually added and stirred
boiling water (about 0.5kg grain to one litre of in after it has been taken off the fire. Like enjera,
water), covered and left to boil. shuuro is preferably eaten with meat and
No preparation is required for Famix, and vegetables or with milk. An alternative addition
the cooking method is simple. is one small tin of tomato paste. Shuuro is
considered 'heavier' than enjera, which means it
Cooking: The main methods of cooking are
is less easily digestible.
boiling and baking in the 'enjera oven' (see below).
Karis and garo take different times to prepare,
The most common daily fare was said to be
but about the same amount of time to cook (garo
enjera. According to the woman interviewed in
the local village, enjera was more commonly might be slightly less). After preparation of the
made in the dry season than the rainy season, grain, it is placed in hot water, brought to the
because it can be eaten without milk, while boil, and covered with a lid. While these were
slmuro should be taken with milk, which is more cooking, the women could go away and do their
readily available in the rainy season. chores, returning before it was ready. For the
study demonstration, 0.5kg of grain was
The enjera. dough is prepared the evening or
afternoon before and left to ferment. The daawa prepared and cooked for each dish. The
is wiped with oil and placed on the fire. When it cooking time was about two hours. Before the
is spitting hot, a cup of the mixture is poured on cooking process was completed for the karis,
to it in a circle, starting at the middle and when all the water was absorbed, salt and spices
spreading it outwards. The diameter is just less (geedo) were ground together and added, and
than the lid of the daawa, which is then put on about 20ml of oil also. Salt and spices are added
and left for about two minutes, after which time to the garo after any excess water has been
the enjera is removed and the next cup poured drained off, and then it is ready to be served.
on the daawa. In the cookery demonstration, Karis can be eaten with meat and vegetables, or
about 1.25kg of enjera flour was mixed with with milk which is poured over it. Garo can be
water to make the runny dough. The cooking of eaten with milk or with oil or mixed with beans,
the enjera took just over one hour. The women but is not eaten with meat and vegetables.

Somali refugees in eastern Ethiopia

Famix was cooked similarly by all women said that they were not cooking lunch that day,
who were asked and observed. They heat some because they had not enough food, whereas in
water, and when it is hot but not boiling they HA this was not the case. It was difficult to verify
add the Famix slowly, sprinkling it into the this, because much of the cooking is done inside
water and mixing all the time. Once it comes to the alawaat, and wood smoke may denote a meal
the boil, it is boiled for 2-5 minutes being prepared, or samosas for sale, or the
approximately (one woman said she boiled for preparation of tea. The use of 'in-between'
15 minutes, because it did not taste good if snacks was rarely mentioned. In KB small
cooked for less; but she was the exception). children were observed nibbling sweet potatoes
Many women were adding sugar because they in the middle of the morning (from the garden
(or more commonly their children) complained of that household), and in HA deep-fried wheat
that it was not sweet enough. Also one or two snacks had been bought from the camp market
women expressed the belief that the sugar was for the children of one of the interviewees.
boiled away during cooking, so that more had to However, for adults, it appeared that tea was the
be added before eating. Some also added more main 'in-between' sustenance, as well as being a
oil and, occasionally, salt. The resulting boorash meal substitute on occasion, according to some
was of a thin, gruel-type consistency often interviewees.
likened by women to 'milk' or 'tea', which is The breakfast was almost invariably said to be
drunk from a cup. It was said that for the plain enjera, cooked and eaten soon after sunrise, the
Famix (not the pre-mix) more water is required. dough having been prepared the evening or
One woman who said she used one part water to
afternoon before. The amounts prepared are
three parts Famix (similar to the ratio recom-
usually for one meal only. It is not stored for
mended by the manufacturers) said she cooked
much later meals, because it becomes bitter,
it thick and gave to the children from a spoon.
though a few women spoke of saving the
The cooking method of Famix is similar to breakfast enjera for lunch and taking it with tea.
that for making shurbad, a porridge-type dish Famix can be a substitute for breakfast, but most
which many referred to having cooked people said they had prepared enjera that
previously when in Somalia. For shurbad, white morning, and almost always they had eaten the
flour or sorghum flour might be used, or tinned enjera. with tea, either taking the tea separately
white oats (which were favoured). The flour or or pouring it over the enjera, and some had
oats were added to hot water, stirred and mixed poured oil on to the enjera prior to eating it. Salt
in with milk and butterfat, possibly sugar. This and onion are sometimes added. One or two
was traditionally made for young children, women admitted eating enjera with spices,
newly delivered women, and the sick; and also onion, and tomato for breakfast. Other dishes
during the Ramadan fasting period. Something like karis andgaro and bur can be taken, but were
similar but made to a much thicker consistency not often mentioned during the study as that
than shurbad was cassiset, which was mostly for day's breakfast.
the newly delivered women.
Women said that, if they had something to
Mostly cooking takes place in a smaller aqal prepare for lunch, they began preparation
(alawaat) next to the home aqal, with very little around 11 a.m. for midday eating, but it could
ventilation. Charcoal is used by some, in a be later than that if the woman had been to
square tin or round clay stove, but it is more
collect firewood first in the morning.
expensive than firewood. It is more convenient,
Shuuro is eaten for lunch or supper, and is
especially in the rainy season and also if cooking
considered a heavy food, even if taken 'dry' —
inside the living aqal, to avoid wood-smoke
pollution. However, particularly in KB, that is without meat and vegetables, or milk. It
firewood was used for cooking, and the was often referred to in connection with supper.
collection of firewood for use and/or sale is a Supper is eaten around sunset, so preparation
major activity for many women. would start around 5.00 p.m. at the latest.
Garo and karis may be taken for lunch or
2.13 Intra-household food distribution supper. In HA, women also mentioned
The most important meals of the day are spaghetti and rice, eaten with a sauce for
breakfast and lunch, so that lunch is usually supper, and admitted to eating meat sometimes
heavier than supper. If something heavy is with the staple dish.
taken at lunch, then supper will be lighter, and Famix was taken mostly as lunch or supper,
vice versa. Most of the women interviewed in KB according to most women, and not as a between-

Acceptability and use of cereal-based foods in refugee camps

meals snack, except sometimes for small some sugar. Use of the bottle was said to be
children. It appeared that people generally common in HA, but the bottles are hidden
would substitute Famix for lunch or supper because they are discouraged at the clinic
rather than for breakfast. It was not considered (according to a camp-based key informant).
equal to wheat flour by most, but in the absence
Commensality: Members of the household try
of flour it was an acceptable second choice as a
to take meals at the same time, but
light food to balance heavy food taken at the
circumstances of camp life make this difficult
other mealtime. One woman (KB) said when
sometimes, because various members of the
she had received 4kg in the general ration some
household are out doing business or chores or
months ago, she cooked it for one week, every
taking leisure time.
day for lunch for everyone: 'It is not good for
lunch, but we were forced by circumstances'. Asked if all members of the household —
Those receiving one ration of pre-mix through men, women and children — eat together, the
the SFP/BFP tended to use the week's ration all women laughed on more than one occasion,
at once or twice, to give to everyone in the presumably because for them it was
household, so that the week's ration lasted only unthinkable that all should take 'from the same
one or two days, or — as another woman said — plate'. The men eat from a separate plate and
it was enough for two meals for four people each preferably in a different space from the women
time. and children. Food is eaten inside the aqal,
either in the entrance section, which is often a
Famix was used for infants as a weaning food, cooking space where charcoal is being used, or
women saying that they gave it with a cup and in the main living/sleeping section. It may be
spoon. However, it was also eaten by all that the older sons have a separate sleeping aqal
members of the household, and did not at all and they would eat there. Plates and pans are
seem to be considered as only a food for washed outside, and refuse is given to chickens
children. Sometimes a small amount was added and goats and/or thrown into a pit. When the
to the wheat flour in preparing the enjera researcher and interpreter were invited to
dough. It was commonly described as 'like lunch in HA by a woman from the Committee,
milk', and it was made to a thin consistency and they were served first, then her husband and
drunk from cups like milk and milky tea. The adult sons, before she sat down to eat.
tea made in the camps is extremely sweet
Thus the woman usually makes sure that all
(reportedly around 0.25kg of sugar for one
are satisfied before she takes food. Some women
Thermos) and fresh milk is preferably added, so
said that sometimes they missed out because the
that tea is more than a drink.
food was not enough. This was most likely to
All members of the household liked to drink happen if unexpected visitors appeared; these
milk, either cold by itself, heated with sugar or would take precedence over the woman and
mixed with tea, or as yogurt, together with karis older daughters, not the small children (male or
or garo or shuuro. For young children it was female). When asked why women did not keep
usually heated a little and sugar added. At the some aside for themselves to prevent missing
time of the study it was cow's milk which was in out, one woman said, 'You cannot do that while
evidence (not camel's), and the price was your child is suffering. You are the responsible
relatively low, because the rains had begun. one cooking, so you have to make sure everyone
However, women said that at the moment cow is satisfied'. One woman said that 'in the Somali
and camel milk were the same price. Goat milk, community to eat alone is not good, and eating
they said, which was even better than cow or together is a sign of peace'.
camel, was also around the same price, but the One woman said that if there was enough
yield of the goat was far less than that of cow or food, and the children took some, went away to
camel. They said children tended to prefer cow play and then returned wanting more, she gave
milk; if camel milk was given to small children, it it to them, but this was applicable only to the
had to be heated a little first or taken in tea. children of the household.
In one small aqal (HA), there was a young As further illustration, one woman in HA who
widow and her 7-month-old baby. She was seemed to be well off (and had certainly been so
breastfeeding and giving reconstituted dried previously in Hargeisa town) explained that in
whole-cream milk (tinned) from a feeding Hargeisa she and her husband and children
bottle. She was adding c.3 teaspoons of milk used to eat seated at the table together, but
powder to 115ml of boiled water, and adding taking from separate plates; this happened

Somali refugees in eastern Ethiopia

when it was just the close family and there was town-dwellers, from both northern and
food in plenty, so she knew there would be southern Somalia, tended to mention first rice
enough. If there was not much food, the women and spaghetti, and 'white flour'. This white
would wait until the others were satisfied; or if flour was wheat flour, but several claimed they
visitors came, the women would retire to did not know from what it was made and were
another space and wait until all had received not familiar with wheat grain before they came
enough food, before they themselves partook. If to the camps. With the white Hour they would
the food was not enough for the women, they make enjera sometimes for breakfast, eating it
would cook again, or buy from the hotel. with oil and tea, or with meat stew (faxfax). This
The same woman went on to explain that, white flour was also made into white bread in
before coming to the camp (in her experience), bakeries; this was bought and eaten for
family members would be sitting down to eat at breakfast by many in the towns. One woman
the same time, because they were living in said the flour was sweeter than that obtained
houses with a spacious compound and good from milling the ration wheat, and the bread
fences, so that the children stayed in the tasted sweet also. Some said they had used
compound. In contrast, she said, in the camp sorghum in the towns, but it was factory-milled
there were no fences, and children were sorghum, unlike whole-grain sorghum, which
running all over the place, and the adults were you have to mill and which is too coarse by itself
out 'running after money', so that people were to be enjera. A number of people talked of
returning home at all different times. mixing beans with sorghum for a main meal.
Among former town-dwellers and the well-off,
2.1.4 Food preferences spaghetti and rice were highly favoured,
A number of women seemed to favour wheat spaghetti probably more so than rice. Both
over sorghum and maize, because they said it commodities were on sale even in KB, but the
was more versatile, in that it could be garo or
price of both was much higher than the other
karis (pounded, boiled, but not milled), enjera,
shuuro or bur (requiring milling). Wheat flour is
Those from a pastoralist or agro-pastoralist
also used by several women to make samosas for
background tended to favour sorghum over the
sale each day in the market areas. Another
reason, which they did not state, but was other grains, and placed great emphasis on
explored in the SCF-UK HFEA, was that in KB, meat, and particularly on milk and butterfat.
at least, for certain periods of the year, the price Cow's milk was generally preferred over
of wheat is higher than that of sorghum and camel's, but they said that children could take
maize, so wheat grain represents potentially either. One reason why cow's might be
more cash available to buy other food, and non- preferred (as explained by the local villager
food items. In fact, asking about preferences in interviewed) is that cow's milk can make yogurt
interview was not easy, because the women and butterfat, but camel's milk cannot. When
made a distinction, understandably, between they sold an animal, explained one woman, they
what they preferred before when they were in would buy rice or wheat, but they would not
Somalia and what they preferred now, because make enjera with the wheat. Rather, with the
now the choice was so limited and heavily wheat she would make shuuro and garo most
dependent on what was given in the 'gift'. They regularly. The rice was expensive, so they would
may have feared that expressing preference buy it only for special occasions.
other than wheat to the researcher might There were some contradictory replies in
prejudice future supply in some way. relation to whether it was pastoralists or urban
In terms of choice between wheat, sorghum, dwellers who had most commonly eaten enjera
and maize, now that they were in the camp most before in Somalia. One woman said that some
said they preferred wheat first, then sorghum, pastoralists did and some did not, that it
then maize. When asked why, in that case, they depended whether they could afford to buy the
were growing so much maize in their gardens, wheat flour. Generally it seemed that enjera was
they replied that it was easier to grow than eaten more by former urban residents, and
sorghum, and that when it was ready the shuuro more often by nomadic households. One
children could take and eat asxasiid. woman from an agro-pastoralist background in
Most interviewees were asked what kinds of the south of Somalia said that they used to grow
food they commonly prepared before they sorghum and maize and preferred sorghum,
became residents in the camps. The former because with it they made shuuro and garo

Acceptability and use of cereal-based foods in refugee camps

mostly, and because they considered it more children refused it, because it was not sweet
nutritious than maize. She said (as did some enough. Asked why they did not reject other
others, though it was disputed by one or two) foods like enjera. on that basis, one woman
that they did not eat enjera then, and that some explained that enjera was not expected to be
of the nomadic peoples did not eat enjera at all, sweet, but boorash was. For this reason, the
that it was mostly town dwellers who did. Here Famix was often served with extra sugar added,
in the camp she was growing maize, however, as was milk. Many said they did not know this
because the children could eat xasiid and also type of boorash before in Somalia, but many had
because the garden was not very large. made shurbad; and those from towns had often
Another woman from a pastoralist used the tinned white oats for this and
background said that she preferred sorghum to perceived them to be sweeter and more
wheat, but sorghum alone cannot be enjera nutritious than the Famix boorash. However, the
(according to local villagers and key informants, Famix boorash, said some, tasted better than
it can), and when the children saw enjera being wheat-flour boorash without milk and sugar.
cooked by others, they asked her to make it for
them. She said there was white and red 2.1.5 Local factors influencing food practices
sorghum, of which she preferred the red, but Understanding of health and nutrition issues:
only because of the colour, as the taste and Women in interviews did not make any
texture were the same. Another woman reference to specific micronutrient deficiencies
expressed preference for the red sorghum and such as scurvy; in a discussion of the scurvy
said that the one supplied as 'gift' last year wasoutbreak of many years ago, they did not seem
good, even though it was not the same as she to distinguish it from the problems of diarrhoeal
had grown herself before in Kismayo. However, disease and wasting and hepatitis, which they
this same woman said that now she was in the remembered from those days.
camp she preferred wheat to sorghum, because When asked about special foods during
she could make more things from it (like pregnancy, women would speak of a 'balanced
samosas, which she was selling in the town). Also diet', by which they meant milk, meat, fruits,
she said that before becoming a refugee there vegetables, eggs, and liver. They said that
were many things she was not familiar with: for pregnant women particularly should eat more
example, she used to wear one set of clothes all meat and vegetables. There seemed to be no
the year when in the rural area, but now she had food avoidances or taboos for women during
got used to being in town and wearing different pregnancy and delivery, but some said that food
sets of clothes. This woman (previously like shuuro with oil was less appetising and not so
easily digested when pregnant, and foods like
pastoralist) said she had seen enjera before in the
towns, but did not know how to cook it until she meat and vegetables were preferable for that
came to the camp, while another woman reason also. Cassisel and shurbad made from
(previously living in Hargeisa) said she had not sorghum (for pastoralists — sometimes from
eaten shuuro before coming to the camp and had white oats for the town dwellers) were prepared
subsequently learned how to make it. customarily for newly delivered women, and so
Famix could also be used for this purpose,
Virtually all of them expressed preference although it is probably not thought of as quite so
for enjera and shuuro over the boorash made from 'nutritious'.
Famix, though one woman said she preferred
Famix because it was easy to prepare. When The feeding of children was of course a
preoccupation with many of the women. They
asked, they said they did not know what it was
often referred to the value of milk for children.
made from, but had an idea it was 'nutritious'
It was considered the most important food item
and contained 'protein', and had accepted it in a
for their nutrition. In terms of commensality,
way as a milk substitute. Several were buying in
children were given priority over older girls and
the market; one woman, when asked why she
adult women. In both camps, a few children
bought Famix and not wheat, said because it was were observed eating mid-morning snacks, but
easy to cook, and she had only enough money the Famix, said the women, was not used for this
for Famix, which was much cheaper than meat, purpose.
implying its perceived value as a source of
protein. The 'camp' situation: The camps are not in any
Some mentioned that, while most people of way 'closed': on the contrary, they are an
all ages liked the Famix, sometimes the small integral part of the regional 'economic and

Somali refugees in eastern Ethiopia

cultural continuum' (1.2). Thus the availability and business ventures were not so accessible nor
of markets containing a wide range of food and so relevant for this study, because those
non-food items influences what foods are involved do not depend on the general ration
prepared, and the scope of food preferences. for nutritional reasons. There appear to be wide
Recent repatriation initiatives too will have had variations in socio-economic status in these
an effect, with many refugees effectively having populations. Those in KB have fewer support
repatriated themselves but having kept a networks, because they do not live in their own
presence also in the camp, and with others now clan areas, unlike the residents of the other camps.
taking up the repatriation package on offer One woman (nomadic background), when
from the agencies. This will influence food asked whether she worked harder in the camp
production, food trading, and food-security than she had in her home place, said it was
strategies of the camp residents. harder in the camp, as she was collecting water
To meet the study proposal criteria, most of and firewood, gardening, making samosas, and
the interviews were conducted in KB, because tending the children. Her older daughters were
of its status as the poorest camp and because, as not helping, because they were already
such, it was the most dependent on the ration. married, and the husband spent his time going
However, HA was also visited, to ensure that the to the town and talking with the other men.
Isaaq clan was included in the study (the most
numerous clan in North-Western Somalia and 2.2 Opportunities for cereal fortification
many of them formerly town-dwellers), in case
food practices and preferences were different 2.2.1 Household-level milling
from those of the KB camp. The higher Household-level milling of cereals is
economic status of the HA residents appeared undertaken by women. The milling process is
to have an impact on types of food regularly pounding, winnowing, and stone grinding by
purchased, but not on choice of milling method, hand. The equipment used is the viooye and tip,
except that rice and pasta, the more high-cost the shiide and the masaf. The mooye and tip are
foods, do not require milling anyway. made from particular trees which do not grow
locally, so the refugees either brought their own
The Famix, on sale in both places, was said to
with them when they came or bought from
originate in HA. This could reflect the higher
people living in the Ogaden (they said the cost
economic status of its residents, who relied less
was around 20-25EB). Most women said that
on the ration food. However, it could equally be
they shared the equipment with their
due to the fact that the population figures on
neighbours: up to ten women a day could use it.
which food distribution is based are more
Mostly if the grain is pounded at home, it is then
inflated in HA, which is visibly smaller than KB.
hand-ground, but one woman reported that
Opportunities for work and income first she pounded the grain at home to clean it,
generation: Many women, particularly in KB, then took it to the commercial mill, because, she
spoke of collecting firewood to sell in the town to said, the commercial mill could not clean it well
buy food. A morning of collecting firewood enough. One woman who was using the mill in
could earn about 4EB. Women were seen HA said she used it most times, but it was not
making mats to sell for covering aqals, and quite fine enough, so she sieved it afterwards at
several were making samosas. This was not home with the masaf. So in some cases the
necessarily something they had done before in milling process is a combination of household
Somalia. Some mentioned men out working, for and mill, either before the grinding or
example one spoke of her husband out afterwards.
collecting large pieces of timber to sell to For several women it seemed that, if they had
building contractors. no money for the commercial mill, the
Credit arrangements were mentioned by alternative was hand-grinding with the stone,
many, often to maintain a small business such as rather than the free camp mill (2.2.2). Some said
samosa sales. Several women had run small they would hand-grind if they were making
businesses before becoming refugees, so skills in shuuro, but take it to the commercial mill for
borrowing and managing credit are wides- enjera. For sorghum and maize when they had
pread. The visible income-generating initiat- it, they were more likely to use the stone than
ives, and those accessible to a visiting research- they would for wheat, although, of the three
er, tended to be those operated by the poorer cereals, maize is considered the most difficult to
refugees. The prolific and large-scale trading hand-mill with the stone.

Acceptability and use of cereal-based foods in refugee camps

Because of the difficulties encountered in they submit to obtain the flour. The operation
camp-level milling and those anticipated in looked speedy and efficient. The cost was
regional-level milling (2.2.2), WFP recently around 200Ssh (0.17EB) for about 1kg.
investigated the possibility of providing hand- There were no overt complaints during the
grinding mills which would be used by groups interview about the milling capacity available in
of women from 5-10 households on a co- and around the camp. Wheat was usually said to
operative basis and/or for income generation. be the preferred staple, even though it is not the
The maximum capacity of these mills would be easiest to hand-mill (sorghum being the easiest,
4kg/hour. The investigations into appropriate then wheat, then maize). The reasons for this
types of mill found that the type available were probably a combination of a desire to give
through a NGO in Addis was not good for the answer they thought would be most likely to
sorghum and was probably not durable influence the food basket in the way they
enough. Alternative sources in Kenya were wanted, the higher market value of wheat over
sought, but the investigation itself was time- other cereals (at least seasonally), and the fact
consuming and expensive, and the procure- that wheat is a little more versatile (suitable for
ment and importation of the large quantity making samosas and bread). The availability of
required would also be expensive, and it was milling facilities was not raised by women as an
unlikely that the supplier could meet the issue or major constraint.
demand. In both camps some women said they were
2.2.2 Camp-level milling too busy to go to the free camp mill, which was
Milling at camp level is available. In all camps at far from their section, and the free camp mill in
least one mill is provided free of charge, but the KB was working only sometimes. In KB, women
maximum daily capacity of each is 50-750kg. tended to say that when it was working it made
UNHCR was responsible for camp-level milling the flour too coarse, while in HA they said it did
until 1996, when WFP took over. WFP pays for not spoil the grain. It may be that those who live
fuel, maintenance, guards, and any other nearest to the mill tend to use it most, but in HA
running costs incurred, but the implementation many were choosing the commercial mill for
of the service is the responsibility of ARRA. reasons which were not clear. Some spoke of
According to one agency key informant, the KB certain camp residents not being eligible; others
mill broke down several weeks ago but was said the quality of the flour was poor, but others
working only intermittently before that, even said it was good. In both camps people said the
though it was relatively new. In other camps not free mills were busy for the first few days after a
included in the study, new mills have been distribution, because that was when the wheat
installed, but they too have functioned below was available. This implies that the free mills do
their capacity, partly because it was several not operate until or unless they have a lot of
weeks before they began operating after the customers at one time, as milling small amounts
new installation (this may have been related to sporadically is not economical in terms of the
opposition from local commercial mill- fuel. However, if it was the case that the mills
operators). During the agency feedback session operated only around distribution time, that
for the study, it was clear that opinions differed does not explain why in both places the small
among key agency staff over how well the mills commercial mills were doing plenty of business
were functioning and the reasons for every day of the researcher's visit, irrespective
inadequate functioning. A monthly report of of when the distribution had taken place.
mill-output is supposed to be produced, but no In KB the operator of the non-functioning
one present had seen such a report recently. A mill said on the researcher's first visit that it was
field-level analysis by implementing and not working that day because the people were
funding agencies is probably required. waiting for the next wheat distribution, which
There are numerous commercial mills in and was already late. He said that if there was at least
around the camps. An attempt by the 25kg to be milled, they could start the machine;
researcher to enter one in HA for the purposes but if it was smaller amounts, they had to wait
of the study was firmly repelled. It was not clear for several people to turn up. However, on a
why the mill operators appeared so hostile to subsequent visit, immediately after a wheat
visitors. In KB the mills were busy on the day of distribution, it was still not in use and there was
the researcher's visit. People hand over their no sign even of prospective customers. This
grain and are given a numbered receipt, which time, he said the mill was not used because it

Somali refugees in eastern Ethiopia

spoiled the grain: it milled too coarsely if less This makes the feasibility of milling regionally
than 50kg was entered. He also said that there extremely problematic, even if there were
was a problem of fuel. A number of women said adequate milling capacity, because milling and
that the mill worked only sometimes, and when fortification would involve off-loading, de-
it did, it did not grind finely enough for bagging, monitoring of the whole process and
samosas, which a number of them were making then re-bagging and re-loading. This would
for sale. Some said it did not make enjera, others require more time, and so adequate advance
that it did not make shuuro. supplies of grain in-country would need to be
In HA one of the free camp mills was visited. guaranteed to avoid serious delays in
It looked in good condition and there was distribution to the camps. It would also increase
evidence of wheat flour which looked quite the cost of internal transportation considerably,
finely ground. The mill was locked up when the as would a similar option of milling in Addis.
researcher arrived. The operator said that there
were no customers because wheat had not been 2.2.4 Blended-food production and fortification'
delivered recently. However, there was a The blended food distributed to this population
is Famix, a product of the Faffa factory in Addis
continuous sound of commercial mills
Ababa, established since the mid-1970s to
operating around the camp, and the women
produce blended fortified food for relief and
interviewed all said that they used the
commercial purposes. For most of that time, it
commercial mills.
was subsidised by government and supported
2.2.3 Regional-level, milling by UNICEF and WHO, among others. Five
The nearest large mill is in Dire Dawa. The years ago, with the change in government in
capacity was quoted variously from 260 Ethiopia, the factory had to plan to become a
MT/month if working only 8-hour days (more profit-making company, with no more subsidies
than double with 24-hour shifts) to as much as or tax exemptions.
l,200MT/month. In any case, even with 24- The maximum capacity was said to be 50-
hour operation, the current monthly require- 70MT/24-hour day (i.e. working 24-hour shifts)
ment of WFP for the Eastern camps for the roasting process, and up to 30MT/24-
(3,500MT/month) would not be met. Even hour day for the extrusion process. In theory
though there are other mills in Dire Dawa, their this represents a maximum capacity of
capacity is less than that of the large mill. WFP lOOMT/day, but this is not being realised. One
has considered milling at the national or reason given by the management is that the next
regional levels in the past, but found it was not stage of the process, involving the sole hammer
feasible with this level of capacity. There is more mill, cannot match that capacity, so actual
capacity in Addis Ababa, but this would increase output will be less. They are not sure how much
the internal transport costs considerably, as the less, as the extruders are still new, and the
grain from the ports to the Eastern camps does commissioning period had just ended. Accord-
not pass through Addis. ing to other sources, a less specific reason is that
Another constraint in addition to mill management capacity at present is weak. There
capacity is that, for the last year at least, WFP has have been problems in meeting the schedule for
never had more than a two-month advance delivery of blended food to WFP, and this
supply of grain in-country for the Eastern resulted in only sporadic distributions in 1997
camps. There is a problem of supply due to at camp level.
donors' inadequate commitments to this During the tour and interviews with the
programme. In recent months, WFP has been management for the study, there seemed to be
functioning in a 'hand to mouth' fashion, some confusion about the vitamin/mineral pre-
borrowing from other organisations who have mix. The management said that they used to
grain available; but loans can be agreed only buy in their own vitamins and minerals
against the shipping information number (i.e. individually and store them in a specially
proof that the WFP supply to repay the loan is ventilated room. However, WFP specified 2-3
already on the high seas). Thus most of the years ago that for its orders the La Roche
grain going to the Eastern camps is trucked vitamin pre-mix should be used. In interview,
straight from the ports (Assab or Djibouti) to the the acting manager said that as this did not
camps; or, if borrowed, it comes from Nazrat (c. include minerals, these were bought separately
one hour east of Addis Ababa) to Dire Dawa. from a La Roche agent and mixed in with the
There is no buffer stock in Dire Dawa or Jijiga. vitamin mix. There appeared to be some

Acceptability and use of cereal-based foods in refugee camps

confusion when trying to ascertain proportions to give advice on how to cook it, or how to
for mixing. However, there is no way in-country distribute within the family — all of which is
that the product can be checked against these important if refugees are to understand the
specifications. purpose of blended food as a strategy for
Another source of fortified blended food is preventing micronutrient deficiencies in the
HCFM (Health Care Food Manufacturers), whole population, as distinct from a therapeutic
established six years ago by a former General nutritional intervention for malnourished
Manager of the Faffa factory. It does use a children. Once the blended food is being
vitamin/mineral pre-mix direct from La Roche, distributed regularly to this population, they
containing all specified requisite items except will be receiving a plain Famix, which is
calcium, which is acquired and mixed supposed to be eaten by all the family, and a
separately (because of its bulk). Unfortunately Famix pre-mix, which is targeted at children
there is no capacity anywhere in Ethiopia to under 5 years. It is unlikely that those distinc-
analyse samples in terms of their micronutrient tions will be made either by the refugees or by
content, so quality control in-country in this the agencies and camp officials.
respect is not feasible. The mixed messages about food preferences,
The capacity of HCFM, which produces particularly with reference to cereal foods, that
'Tenamix', a blended food like Famix, is at were received from interviewees and key
present 20MT/24-hour day. However, capacity informants reflect the factors which act to shape
will soon increase to 35MT/24-hour day, with preference and shifts in preference. They
the acquisition of one more hammer mill and include resale value, perception of nutritional
roaster. Currently they have no extruder, but a value, and recognition of, familiarity with, and
new plant is being planned in another part of confidence in cooking the food item. Blended
town. food in this context appears to have had
WFP's most recent order of C.800MT has intrinsically no features which would lead to its
been split between the two companies, 50:50. rejection by refugees, except if it were perceived
WFP's monthly requirement for the Eastern as a substitute for an item (or a part-substitute
camps is 260MT for the general ration and for an item) which they would value more.
90MT for the TFP/SFP/BFP, making a monthly The political complexity of this refugee
total of 350MT. situation inhibited the study in a number of
Specifications of Famix and Tenamix are ways. There was the wider political context of
included in the appendices of the main report. Somali region 5, which is considered by some to
have potentially secessionist aspirations, within
the relatively new Ethiopian federal state,
Conclusion meanwhile being supported (strongly if
indirectly) by the food aid to the refugee camps.
Use and acceptability of blended food as At the camp level, there was mutual suspicion
provided in the general ration was in one sense and resentment between the remaining aid
difficult to ascertain, since it had been supplied agencies and camp residents. There is a long
only twice. The camp residents were used to and complex history concerning issues of
receiving it only through the selective feeding registration, repatriation, and resettlement of
programmes. In 1996 blended food was returnees. Many camp residents seem to
introduced into the planned ration, but not perceive the agencies as neglectful of their
primarily to address micronutrient deficiencies. needs and offering inadequate support, while
One of the main reasons for its introduction was the agencies perceive the refugees as milking
nutritional, as part of a revised food basket the system and enjoying a prolonged
which aimed to meet more of the nutritional dependency which for many is no longer
needs of the camp population, given the appropriate, given that return to northern
apparently poor nutritional status in some of Somalia is now possible.
the camps in mid-1996. It was also part of the There was particularly a considerable sense
plan to diversify the food basket to reduce the of grievance related to the amounts in the food
amount of grain ration being sold to obtain basket and its unpredictability from month to
other items. There seems to have been little or month, although, as Farah (1994b) points out,
no public information around that time to in terms of food security, compared with
distinguish blended food from the staple grain, unreliable and supplementary activities such as

Somali refugees in eastern Ethiopia

petty trading, selling firewood etc., 'its reliability convenience, and its use does not imply any
affords it special significance'. It has been opinion about the legal status of the territory.
supplied to this population now for nine years. The term 'Somali' is used to denote those of
The many commercial and business-vested ethnic Somali origin, whether their residence
interests in and around the camps in the region is Ethiopia or Somalia.
are a major factor in all programming. The 2 The GOE/UNHCR 'cross-mandate approach',
existence of these powerful interests probably c.l 992, was an attempt to combine voluntary
meant that three issues arising from the field repatriation of refugees, resettlement of return-
work were very difficult to elucidate at the time, ees, and development aid to the Ogaden
namely the monitoring of distribution and region. It had only limited success, and the
refugees' knowledge of entitlement; the sale of influx of refugees in 1994 disrupted plans.
Famix in large quantities; and the poor 3 Afilmfor LWT, made in early 1990, includes
performance of the free camp mills in contrast a short discussion about the possible addition
to the functioning of the commercial mills. of vitamin C to the water supply which was
tankered in to the camps and then stored in
and distributed from rubber-lined tanks.
Notes However, this was not done. The reasons
were not make clear in the film and are not
1 'Somaliland' in this text is used to denote the mentioned in secondary sources consulted
borders of the former British Somaliland and for this study.
of the subsequent North-West region of 4 At the time of the study, 1EB = 1,200 Somali
independent Somalia. It is used in the text for Shillings (and 500 new Somaliland Shillings).

Regions -c
7-11.SEPAR (Southern Ethiopia
People's Administrative Region)
12. Cambela
13. Harar (Metropolitan)
14. Addis Ababa (Metropolitan)
— National border
Regional boundary
Zonal boundary
® Regional capital

Note: The regional status of
Dire Dawa remains contested
Appendix 3(b): Location of camps, Somali Region

km 0 50 100 150

DarwonajiA\j- e f e r i b e r

Jijiga \ • Hargelsa

Camaboker^ *

Refugee camps

Main towns

International boundaries



Date of Establishment APRIL 1988 May 1988 May 1988 May 1988 1991 1991 1991 1991-1992

Estimated population 58,695 24,865 45,000 31,932 43,006 46,379 15,711 10238

Under 5 population 5017 4790 6298 5668 3322 4232 6298 5017
Malnutrition < 80% W'H 18.6% 14.5% 15.4% 16.8% 20.2% 16.2 18.0% 19.8% o

Malnutrition < 70% YV/H 3.3% 2.0% 1.5% 1.7% 3.7% 2.0% 2.4% 3.3%

Crude Mortality Rale 0.5/1000/MO 0.25/1000/MO Not reported 0.1/-1000/MO 0.18/1000/MO 0.4/1000/MO 0.3/lOOO/MO .

< 5 Mortality Rate 0.2/1000/MO 0.6/1000/MO Not reported 3.3/-1000/MO 0.9/1000/MO 1.7/1000/MO 0.32/1000/MO 1.7/1000/MO
SFP Number (Benef) <S 1256 264 479 832 2834 2295 138 937

TFP Number (Benef) 50 13 52 10 65 23 10 HI

Distant KM Nearest Town Hartisheikh 3-8 Km 1 Km 1 Km 1 Km 1 Km Jijiga 90 30 Km 500 M

Markets liarlisheik-Taiwan Good Daror - Taiwan Good Good OK SmaU Good 3
CD Flourishing Flourishing •a



Water Supply Litre 2.5 • 3.0 Liters/P/D Birka Earth dam Birka Earth dam Birka Earth dam 2 - 3 Lit/P/D 2-3 Lit/P/D 2 - 3 L/P/D 2 - 3 Lit/P/D

Food Basket (Wheat SOOg 1900 KCAL/P/D as planned but due irregular delivery of oil actual received about 1700 Kcal. Wheat grain sold to diversify the food basket to buy milk,meal and sugar.
OU 30g, salt)

Soap Last distribution In January 1996 Poor Poor

Grinding mill Poor in all camps Poor Poor

Shelter and Blankets not satisfactory Last distribtlon In 1991

Cooking Fuel Self provision by Refugees Fire/Wood collection, Long Distance walk

EDP Location Distance llarlisheik Rabasso Daror Camaboker Darivonaji Tefcriber Aisha Kebribeyali

Observations: <5 Mortality rate for all camps 1995 = 10.4/1000/yr (National Eth = 162/1000/yr)
Oil = Nov.l995-April 1996 distributed only three times, Crude Mortality rate for all camps 1995 = 3.3/1000/yr (National Eth = 18.1/1000/yr)
supply is irregular and the Kcal/Person/Day was = 1750-2020
Taken from 'Ethiopia: Joint WFP/UNHCR Donor Food-Aid Assessment Mission, June 1996'
Appendix 3(d): Refugee numbers


AUGUST'89 167,000 53,000 65,000 52,000 24,000 10,000 371,000'

JUNE'90 380,000'

LATE '91 500,000'
(650,000 cards
in circulation)

MARCH" 70,000 18,000 60,000 40,000 7,000 5,000 5,000 ie 205,000 excl.
1994 Aisha

SEPT '94 185,000
NOVEMBER '94 185,000 + new
influx of 90,000
= 275,000

END 199S 275,189
FAM JUNE 1996 working figure
NOV 1996 58,708 10,299 46,379 43,006 31,932 45,011 24,865 15,282 275,482

AUG-OCT 21,175 < 5 years
1996 registered for
BFP. At 20% of
total population
this = estimated
total population
of 105,000.

1 Toole et al (1990)

2 van Brabant: 1994

3 Farah: 1994b

Acceptability and use of cereal-based foods in refugee camps

Appendix 3(e): Some market prices at the time of study

Commodity Cost/ Unit(i) Jijiga town Kebribeyah town Hartisheikh
market market
Wheat EB/kg 2.50 1.25-1.50 (ii) 2.00
Sorghum EB/kg 1.20 2.00
Maize EB/kg 1.70 1.25 2.00
Famix (premix) EB/kg 1.80-2.20 (iii) 2.00
Famix (plain) EB/kg 1.00 1.50

Spaghetti EB/kg 6.00
Rice EB/kg 4.00

Milk - cow EB/litre 2.00
Sugar EB/kg 4.00

Firewood EB/bundle small 10.00
Charcoal EB/sack (iv) 6.00 - 8.00

i In the vicinity of the camps, the main currency is the Somali shilling. At the time of this study the
'exchange rate' was Somali shilling :Ethiopian Birr (EB) 1,200:1 and in Hartisheikh, where the
Somaliland shilling is used, it was around 500:1 (Somalitand Shilling :Ethiopian Birr).
ii The price range depends on the type and quality, and how clean it is.
iii The price range depends on how much sugar has been added.
iv A sack is a 50kg sack (by volume not weight).

Appendix 3(f): Chronology of malnutrition
and micronutrient deficiencies

General ration Blended food Malnutrition rates Vitamin C Mortality Comment
(bf) in general (WFH/L) deficiency rates
ration (clinical scurvy)

Sept- wheatflour Faffa - irregular nutritional status on no cases Dec: Wet feeding SFP
Dec oil arrival satisfactory identified CMR: and TFP
1988 <3/l,000/mth

1989 supply erratic Faffa/CSB - 23.6% in March l%-2% Feb:- CH prog (SCF-UK)
irregular 22.9% in May prevalence CMR: hepatitis

1990 supply erratic irregular 4-8%(HA,HBonly*) in TFP and SFP,
c.30 cases/month

1991 supply erratic irregular c.5% HA.HB new camps of TF,
18-24%KB,TF,DW DW, KB (•)

1992 supply 33%<official Faffa C.1O%HA,HB
requirement. No bf discontinued in August: - 6-9% in
included after July July KB,TF,DW

1993 9.7% (average overall
excl. Aware camps)

1994 16% (average overall invalidation Sept.
excl. Aware camps in new influx Oct.

1995 10% -13% CMR:
(all camps) 3.3/1,000/yr
1996 FAM recs:- FAM recs:- 15.2%-21.1% (May) BFP introduced
cereal 400gm Famix 30gm July
oil 2Sgm
sugar 20gm
salt Sgm
Famix 30gm

up to wheat grain Famix no survey conducted anecdotal reports
June oil distributed x 2 since mid 1996. of occasional
1997 only cases

(•) TF = Teferi Ber, DW = Derwonaji, KB = Kebribeyah, HA = Hartisheikh A, HB = Hartisheikh B

Acceptability and use of cereal-based foods in refugee camps

Appendix 3(g): list of key informants and clan/position
in household of interviewees

Key informants: agencies/other Ms Halimo Hussein, Women and Children's
Officer, Jijiga
ARRA Ato Mesfen, Senior Programme Assistant, Jijiga
Ato Ayelew, Director, Addis Ababa Abdi Aweil, Technical Officer, Jijiga
Dr Berhanu Debaba, Medical Co-ordinator,
Addis Ababa WFP
Dr Berhanu Woldesenbet, Mr Justin Bagirishya, Deputy Country
Zonal Health Co-ordinator, Jijiga Director, Addis Ababa
Ato Kahsay Gebremariam Desta, Eastern Ms Daniela Owens, Refugee Co-ordinator,
Region Refugee Programme Co-ordinator Addis Ababa
Ms Lynne Miller, Jijiga
CARE Jijiga: Programme Manager

Ethiopian Health and Nutrition Research Key informants: camp level
Institute: Ms Abeba Gobezie, Deputy Director,
Addis Ababa Kebribeyah Camp
Mohamed Bedele, Camp Co-ordinator
Faffa Food Factory, Addis Ababa Ato Daniel, Health Co-ordinator
Tsigereda Berhane, Acting General Manager Ahmed Haji Abdulkadir, Camp counsellor
Ibaado, Women's Committee
HCFM, Addis Ababa
Ato Belete Beyene, General Manager Harlisheikh A Camp
Ato Amaha Haile Mariam, Camp Co-ordinator
Oxfam GB Dr Bekele Negussie, Health Co-ordinator
Yezichalem Kassa, Health Adviser, Safia Abdi Ahmed, Women's Committee
Addis Ababa
Ato Aregawi Hagos, Deputy Country Rep. Household interviewees: 20 refugee
Addis Ababa households + one local household
Mr Ahmed Abdurahman, (hhh = head of household)
Programme Manager, Jijiga
Place in household:
SCF-UK Female hhh 5
Ms Rachel Lambert, Deputy Field Director, Wife of hhh 14
Addis Ababa Senior wife of hhh 1
Ato Abebaw Zelleke, Regional Manager, Jijiga Widow related to hhh 1
Ato Wondwessen Tsegaye, Nutrition Adviser,
J'j'ga Ethnic groups:
Hawiye 3
UNHCR Ogadeni 6
Ms Allison Oman, associate nutritionist, Hard 3
UNHCR-RLO Addis Ababa Marehan 3
Dr J.Tabayi, health co-ordinator, UNHCR- Darod 1
RLO Addis Ababa Isaaq 4
Mr Abdulrauf Farooqi, Head of Sub-Office, Ogadeni married to Isaaq 1

4 Case-study: Burundian refugees in Tanzania
(August/September 1997)

1 The context boat at Kigoma town during that period (DMA
The only refugees in Kigoma region up to the
1.1 Burundian refugees in Tanzania end of 1996 were Burundian. At the end of
There have been two previous influxes of 1996, Congolese' refugees were accommodated
Burundians into Tanzania, the first in 1972, the in camps in Kasulu district and Kigoma town,
second in October 1993, following the killing of but none in Kibondo district. At the time of this
the President, Melchior Ndadaye. The violence study, there were no longer Rwandan refugee
which followed the killing led to an exodus of camps in Kagera region, following the
around 750,000 Burundian refugees to Rwandan repatriation at the end of 1996.
Rwanda, Tanzania, and Zaire. Most of these The problems of Burundian refugees in
subsequently returned home as Rwandans Tanzania should be seen in the wider context of
arrived in Tanzania and Zaire following the the Great Lakes Region. The genocide in
genocide which began in Rwanda in April 1994. Rwanda in 1994 led to immense population
In particular the Burundian refugees in movements, with more than two million
Rwanda returned to Burundi after May 1994. Rwandans crossing international borders to
The latest major influx reached a peak in seek refuge in eastern Zaire, Tanzania and
November 1996, and part of that influx Burundi, when there were already some
includes the population discussed in this case- 750,000 Burundian refugees in those three
study. It was preceded by smaller influxes countries.
throughout the earlier part of 1996 and Thus the two refugee-affected regions of
following the coup by Major Buyoya in July, but western Tanzania now are Kagera region
the mass influx of November was a result of an (Ngara and Karagwe districts) and Kigoma
escalation of conflict within Burundi at the same region (Kibondo and Kasulu districts). This
time as other critical developments in northern
field study was based in Muyovosi, which is one
and southern Kivu in Zaire (fighting between
camp of Burundians in Kasulu district (Kigoma
the Alliance of Democratic Forces for the
region). It was formally established in
Liberation of Congo-Zaire (ADFL) and Zairean
September 1996, and received the bulk of its
Army) and the subsequent repatriation of the
Rwandan refugees by the Tanzanian residents in November 1996.
government. These 1996 refugees (the 'new Proximity of towns and markets/farming
caseload') went to new camps in Kibondo patterns: The camps in Kasulu and Kibondo
district (Mtendeli and Nduta) and Kasulu districts are all located on either side of the main
district (Muyovosi and Mtabila extension); some road linking Kasulu, Kibondo and Ngara
also joined existing camps. The influx subsided, towns. Along almost its entire length, the road
but continued through the early months of runs between 20 to 40km away from the border
1997. with Burundi (Appendix 4(a)).
Many of the new arrivals had a history of There is a market in Muyovosi camp and one
internal displacement within Burundi for in Mtabila. Local villages a few kilometres away
months preceding their crossing the border. have small markets on some days. The nearest
Small numbers of Burundians were also town market to Muyovosi is Kasulu town, 42km
entering Tanzania from Zaire from November away.
1996 to April 1997. It is estimated that 9,700 The roads in the region are very poor, and
Burundians and 2,200 Rwandans, who had vehicles get damaged easily, so transporters and
been refugees in camps near Uvira, arrived by drivers are sometimes reluctant to take the risk

Acceptability and use of cereal-based foods in refugee camps

involved. Some repairs have been undertaken, originally envisaged to be made available for
but there are many roads which will be cultivation by the Mtabila refugees was used
extremely difficult to pass in the rainy season. instead to settle new refugees from Burundi
WFP is building up buffer stocks in the (UNHCR Thiadens 1996). The site formerly
extended delivery points (EDPs), to minimise was not under cultivation, but was an area of
disruption to the delivery of the food thick 'bush' which had to be cleared.
requirements during that time. A limit is now applied to all camps, whereby
Kigoma Region is considered one of the refugees should not go outside the camp more
poorest regions in Tanzania. The districts of than 4km in any direction. This is not easy to
Kasulu and Kibondo are suitable for the enforce, but affects the refugees' ability to
cultivation of a wide variety of food and cash complement their food rations with income-
crops, including maize, cassava, beans, generating activity and food production.
groundnuts, tobacco and cotton. In the higher Environmental degradation had been a major
areas with higher rainfall, the soils are suitable concern in Kagera Region to the north, but not
for coffee, bananas, pineapples, maize, beans in Kigoma Region up to 1996. However the new
and cassava. These features are similar to the influx during that year led to concerns and to
agricultural situation in Burundi. environment-protection initiatives. In Muyo-
vosi, attempts to control firewood collection, for
Physical features of the camp: In 1993 a joint
example, include the designation of a forest
task force of UNHCR and the government of
approximately 2km outside Muyovosi, where
Tanzania, working on site-selection criteria for
forest guards are stationed, supervising the
camps and settlements, stipulated that the
refugees to do selective cutting to preserve the
usable space should be enough for 'farming
activities', an area of around 3 acres per family green trees.
where feasible. This criterion was established The camp is situated across a stretch of low-
against the prevailing background of lying hills. Refugees build houses of mud walls,
government policy, which aimed at the some of mud bricks, with wooden supports and
establishment of agricultural-based settlements grass roofs with UNHCR plastic. They have a
of Burundi refugees, but not settlements for partition between the entry area and the
long-term integration, only 'semi-permanent sleeping area. The allocated plot size allows for
settlements' for the short and medium term. only a very small vegetable garden area, in
However by 1996, for the new sites being visible contrast to the larger plot sizes of
established, including Muyovosi, the planning adjoining Mtabila settlement.
for plot allocation allowed for family plots of Humanitarian aid: The number of NGOs
only 20 x 25 metres, and agricultural land was involved in implementation of the
not envisaged. The UNHCR agricultural humanitarian response in western Tanzania has
adviser commented in a report of September declined dramatically since the repatriation of
1996, 'These sites will thus be developed as the Rwandans at the end of 1996. Most of the
camps, where refugees will continue to be fully activity until then was in Kagera Region to the
dependent on care and maintenance assistance,
north, where the number declined from 32 to
unless policies change and farming land is made
eight (five of which were involved in remaining
available' (UNHCR Thiadens 1996).
refugee 'care and maintenance' programmes,
In fact the overall size of Muyovosi camp is three of them involved in support to refugee
4.5 sq km, with 1.5 sq km actually usable for affected areas). In Kigoma Region, in contrast,
habitation, giving currently a plot size of around there were only around 30,000 'old caseload'
15 square metres. This is in contrast to the Burundians until the influxes of 1996. Here the
established older settlements such as numbers have increased and the focus of aid
Kanembwa in Kirondo district, for example, agencies has shifted to this region, with many
where the plot size is around 40 x 50 metres.2 national and international aid-agency staff re-
Even in the case of the older 'settlements', legal locating. However, with the prospect of
documents stipulating the land-use rights of repatriation of Congolese and the registration
refugees are non-existent, with the local in July 1997, budgets are being revised
authorities tolerating cultivation by refugees, substantially.
but contradictory policies within the regions
dictating how much they are allowed to Repatriation: The Rwandese refugees in the
cultivate. In the case of Muyovosi, the land adjoining Kagera Region repatriated at the end

Burundian refugees in Tanzania

of 1996. At the time of the field study, plans for were not being admitted into Muyovosi camp,
repatriation of the Congolese were under way, which was officially full and so closed to
and small numbers were returning to DRC. newcomers. They were being housed in Mtabila
However, because of the continuing precarious extension.
situation in Burundi, there are currently no
plans for repatriation of the Burundian Ethnic group/composition: Many of the 'new
refugees. caseload' refugees came from Ruyigi and
Rutana districts in Burundi, which border
The Tanzanian government maintains a Tanzania, but some were from more distant
policy of a 'closed border with a human face'. In districts, including Gitega, Karuzi, Bujumbura,
Kagera Region this leaves, at the time of the and even Bururi and Cibitoke. They are
study, around 100,000 Burundians; and exclusively of Hutu ethnicity. However,
around 131,000 Burundians and 73,000 according to at least one commentator, while all
Congolese in Kigoma Region. the refugees are of Hutu ethnicity, this has been,
Security: At the beginning of the field study, used in the past by educated refugees in
there were tensions along the Burundi- Tanzania as a shield to hide 'a divisive refugee
Tanzania border, with troops from the society from view' (Sommers 1995). These
respective countries reported to be increasing divisions are based on educational levels and
their presence. In Muyovosi, after about a week political allegiances, and to some extent are
the tension had subsided somewhat, following a related to place of origin, with particular
visit by the Tanzania Chief of Defence to the distinctions being made between lowlanders
refugee camps to discuss refugees' security and highlanders. Therefore 'for most Burundi
concerns and to reassure them about the refugees, public silence is the safest strategy for
situation. At the same time, a decision was made survival. The lives of the poor and
to increase the contingent of police and security disenfranchised are often tenuous, and most
guards within the camp, to improve the internal Burundi refugees believe that voicing their
security situation. In previous months there views publicly could not help them' (Sommers
had been several murders of refugees by 1995).
refugees, and a number of violent incidents. In
this milieu any strangers, including obvious Religion: Most of the refugees are Christians of
foreigners, are viewed with some suspicion, so various denominations, Catholics, Baptists,
time was taken before beginning the study in Pentecostals, and others, including 'fundament-
the camp to introduce and explain the project to alist' Christian sects.
all parties concerned. The tension in the camps Language: The refugees all speak Kirundi,
eased during the field study as the refugees which is similar to Kinyarwanda (Rwanda) and
were reassured that border tensions would not the local Tanzanian language of Kiha. In
impinge on the camps. addition, many speak Kiswahili, and a smaller
number, mostly those who have been educated
1.2 The refugee camp population and worked in the professions, speak French.
Numbers: Following a registration of refugees According to Sommers (1995), only a small
in Kigoma region by UNHCRin July 1997, the percentage of women refugees in Tanzania
population of Muyovosi was determined to be speak Swahili, 'thus accentuating the gender-
27,750, which was a 32% reduction in the official specific perspective that officials receive'.
statistic used as a working figure until then
(Appendix 4(c)). A very small proportion of the Literacy: No survey had been done on the
refugees came from Zaire, having been refugees literacy rate. It was assumed to be low, especially
there for periods varying from a few months to among women, and adult literacy classes had
four years (i.e. since October 1993). been started in some of the zones. In Burundi
the adult literacy rate in 1990, according to the
During 1997 the situation was quite fluid,
WHO World Health Report of 1995, was 50%
because of a continuing small influx of
(both sexes), 40% (females).
newcomers and 'recyclers' claiming to be new
arrivals; they had to wait in a transit area (in Occupations: According to the WHO World
Mtabila extension) while their claim was verified Health Report of 1995, in 1992 only 6% of the
by UNHCR and the Ministry of Home Affairs population of Burundi lived in urban areas (for
(MHA). Newcomers from Burundi were still purposes of comparison, it was 22% in Tanzania
arriving during the course of the study, but they (WHO 1990)). The UNHCR camp summary

Acceptability and use of cereal-based foods in refugee camps

identifies the background of the refugees of divided into 11 zones, which were 'cleared' from
Muyovosi as 'peasants, teachers, students, the forest area of Muyovosi site, as the refugees
Government civil servants'. There is a continued to arrive, so that allocation of
comparatively small proportion of artisans, refugees to zones was done on a largely chrono-
tailors and barbers among the Muyovosi logical basis, i.e. according to which zones had
population. These are reportedly more vacant cleared plots for the latest newcomers.
apparent among the Burundian refugees in Each zone has a zone leader, elected by the
Kibondo district (Halepote 1997) people of that zone. The Ministry of Home
A number of refugees are employed in the Affairs (MHA) and UNHCR have established
camp by the agencies, on contracts or as casual guidelines on how this should be done. All in the
labourers. There are some small businesses, for zone should be present, and the candidates
which refugees should obtain permits from the should be identified by colours. Each person
MHA. They include shops, restaurants, bars, then puts a mark on a paper with the different
and the mills. colours represented, in a secret vote.
However, of those arriving in late 1996 and The 11 leaders were all men at the time of this
during 1997, the majority came from rural study, and of their assistants all but one were
areas, particularly the border areas, where food men. There is a Tanzanian government
supplies were affected by the sanctions placed representative and a UNHCR Field Officer also
on Burundi by the international community. present in the camp. Each week, the zone
Political activists and leaders came later, leaders and camp authorities gather in a formal
especially after the killing of a Bishop in meeting to discuss operational issues in the
Bujumbura, fearing reprisals. With only a small camp. NGO implementing partners are also
proportion of artisans, the population of present at this meeting. Similarly the Food
Muyovosi is predominantly rural and a 'very Committee meeting which precedes each
poor caseload' (Halepote 1997). Burundi fortnightly general distribution involves MHA,
refugees are relatively good farmers: skilled, UN, NGO, and refugee leaders. Men do not
accept the idea of voting for women as leaders,
because of their experience of small land-
on the grounds that they are not educated.
holdings in Burundi, in applying intensified
Women are afraid to speak or ask questions in
farming practices. They mostly apply the same
front of men, and there is a feeling of unease in
farming system asTanzanian farmers: 'Climatic
mixed meetings (key informants and personal
conditions allow for two cropping seasons ... the
observation). This was reflected during some of
prevailing farming system in Kigoma Region is the interviews for this study, when the arrival of
mostly low input traditional subsistence rain- the husband, or the obvious presence of men
fed farming of mainly food-crops (maize, beans outside the hut who were listening in, tended to
and cassava). Land preparation is done inhibit the responses of the women.
manually, using the hoe; seeds are mostly local There are various political allegiances within
varieties and the use of fertilisers and pesticides the camps, which creates a degree of suspicion
is not common. The average cultivated area per and hostility between the refugees and a
family is less than 2 hectares' (UNHCR wariness of foreigners. In Muyovosi there have
Thiadens 1996). been seven murders of refugees by refugees in
Age and gender composition: See Appendix recent months.
4(b). This table (produced by a UNHCR 1.3 History of malnutrition and
electronic data-processing team) indicates a micronutrient deficiencies
higher number of adult men than adult women;
also that children under the age of 5 years The following section begins with a brief
represent 17% of the population, and all synthesis of the history of malnutrition in the
children up to the age of 18 years 56%. camps in Tanzania from the beginning of the
influx of refugees from Rwanda and Burundi in
Leadership structure: Muyovosi camp is 1993 up to 1997. After that, the specific history
managed by the International Federation of the of nutritional status and interventions in the
Red Cross and Red Crescent Societies (IFRC) camp of Muyovosi is traced.
and the Tanzania Red Cross society (TRCS),
TRCS being the implementing agency; the 1.3.1 The camps in Tanzania, 1993-1997
camp manager and management team are 1993: This year saw the largest exodus of
employed by this agency. The refugees are Burundians to Tanzania of recent years. In

Burundian refugees in Tanzania

October and November, approximately of angular stomatitis, particularly among
325,000 had moved into Tanzania following the schoolchildren. The prevalence of angular
coup d'etat and subsequent inter-ethnic violence. stomatitis was between 10% and 15% overall.
Malnutrition and mortality rates had risen One reason proposed was that up to the end of
substantially above normal levels. 1995 the blended food ration had been
50gm/ppd, but this had been reduced to
1994: On 28 April, an estimated 170,000
30gm/ppd from January 1996 onwards. The
Rwandans crossed the border into Tanzania
prevalence of angular stomatitis in the local
within a 24-hour period. The unprecedented
population at that time was not ascertained.
scale and rapidity of the influx surprised most
Nutritional status otherwise was satisfactory at
agencies and observers. The refugees arrived
the time of the study.
with a reasonable nutritional status. Adequate
levels of nutritional status were maintained up FAM October 1996: A food assessment mission
to the time of their repatriation at the end of (FAM) was conducted around this time for the
1996, and at least one study showed a entire Great Lakes region, but several specific
considerably more favourable nutritional status recommendations were not implemented,
in the camps than in at least some of the because of the dramatic change in the situation
surrounding villages (Joyce-Jenkins et al 1996). following the sudden and rapid repatriation of
Rwandans at the end of 1996. The mission was
1995: A report in early 1996 by the UNHCR
undertaken just before this event, at a time of
senior nutritionist summarises the situation
great political uncertainty, arising from the
during 1995 in the Great Lakes region: 'Full
sudden outbreak of military hostilities in South
ration of 1,900 Kcal was rarely achieved
Kivu between the Banymulenge and the
at the start of 1995. Since the 1995 October
Food Assessment, average Kcal remained Zairean military. There was also unrest in North
approximately 1,900 in Tanzania... The rates of Kivu. The FAM recognised that certain
malnutrition among children under five in most planning assumptions would have to be
of the camps remain below 5% of the median adapted to rapidly changing realities, but stated
reference for weight and height and the Crude emphatically, 'The need to maintain the
Mortality rates below 1/10,000/day' (Bhatia regional character of the food aid programme,
1996). and the principle that food will follow those in
Up to the end of 1995, the CSB ration was need, must therefore be stressed'.
50gm/ppd, but from January 1996 it was On the use of the ration food, the mission
reduced to 30gm. The ration was therefore as stated that 'only a small amount of the WFP-
below, and remained so up to and including the supplied food was to be found on the market
period of this study (note the relatively high and trade remains within acceptable limits. The
proportion of the pulses ration): proportion of WFP food being exchanged is
estimated to be between 15 and 30%. This
350gm/400gm meal/grain exchange of commodities is undertaken in
120gm pulses order to diversify the diet and obtain
20gm oil commodities more compatible with the
30gm CSB traditional eating habits'(WFP/UNHCR
5gm salt 1996b).
1996-1997: During 1996, as a result of ongoing On the composition of the food basket, the
conflict in Burundi, approximately 58,000 mission recommended that 'Considering the
registered new arrivals were settled in the preferences of the refugees and also in an
Kigoma Region of Tanzania. WFP maintained a attempt to save energy in food preparation, the
high level of food deliveries nevertheless, by target of maize flour versus maize grain should
borrowing food from contingency stocks and be 50% for each commodity'(WFP/UNHCR
EU Regional Food Security stocks and by 1996b). This balance would allow for sale of
effective programming of the regional pipeline grain, which has a higher market value than
at WFP headquarters. flour. It also allowed for some brewing. The
A survey of vitamin B9 deficiency was districts of Kigoma were given priority for the
conducted in September 1996 in Ngara district maize meal at the end of 1996, however,
refugee camps (Rwandan), after the UNHCR because of new arrivals into camps where there
nutritionist and UNICEF programme co- was insufficient camp-level milling capacity at
ordinator noted an apparently high prevalence that time.

Acceptability and use of cereal-based foods in refugee camps

The malnutrition rates in Tanzanian camps Kibondo district and in Mtabila extension
during this year are indicated by the (adjacent to Muyovosi); undertaken by a team
August/September nutrition surveys, which from the Tanzania Food and Nutrition Centre,
gave a global malnutrition rate ranging between Dar es Salaam (TFNC), it was being conducted
0.3% and 12.5%, and an average severe at the same time as this field study. The TFNC
malnutrition below 2% (0.1-1.8%). 'There were study included clinical assessment, but also
two camps with rates above 5% which were information on household food use. Provisional
mainly due to the new arrivals in these locations. findings suggested that the prevalence of
The rates are below the standard rates of the anaemia3 among children and pregnant women
local population and can be classified as was high (76.4% in children <5 years, 50% in
satisfactory' (WFP/UNHCR 1996b). pregnant women) (UN1CEF/TFNC 1997).
With reference to scurvy, an investigation
1997: According to WFP Kigoma, there were no
conducted earlier in the year in Tanzania into
major gaps in the provision of the general-
the stability of vitamin C in the blended food
ration food basket i.e. no more than two weeks,
supplied to refugees from the USA concluded,
except at the beginning of 1997 for two months,
'The amount of vitamin C provided by CSB or
when there were problems with the cereal and
WSB, containing 40mg/100g of vitamin C at the
blended-food deliveries, not because of any
point of consumption, when provided at a
overall problem with availability of food but
ration of 30 grams of CSB or WSB per day,
because of problems of transportation.
would be 3.6 mg/day given a 30% cooking
More oedema has been reported in Kigoma
retention' (SUSTAIN 1997).
Region among the more newly arrived
A re-registration was conducted in Kigoma
Burundians than was observed in Ngara district
Region in late July. The reduction in
among the Rwandans and Burundians who
population figures was dramatic, at 34% overall
came in 1994. A number of reasons for this were
(Appendix 4(b)). The next FAM is scheduled for
suggested by key informants, notably that the
October 1997.
more recently arrived Burundians were in a
poor nutritional state, having been displaced 1.3.2 Muyovosi: nutritional status
inside Burundi for many months. Their staple The refugees' nutritional status on arrival
food is plantains and cassava, which could be a during the last months of 1996 was poor. Many
contributing factor. A contrast was made of them had spent several weeks hiding in the
between the Ngara camps and the Kigoma border areas in Burundi before arriving in the
camps in the timely uptake of available services, camps. Malaria has been identified as a major
the newer caseload of refugees tending to wait problem in all the camps, some worse than
longer before seeking help from the selective others. One reason suggested has been that the
feeding programmes available. Also it was the Burundians are more susceptible to malaria
case that oedema had been somewhat over- because they come from highlands where it is
reported by being misdiagnosed by less not endemic, as it is in Kigoma Region. Another
experienced nutrition staff. However, it is also source refers to a level of chloroquine resistance
highly relevant that in Burundi during this (Genaille 1997). High rates were reported
year, alongside high levels of internal earlier this year from Mtabila camp also, which
displacement and consequent food shortages, adjoins Muyovosi and houses some new
high rates of oedema are being recorded arrivals.
compared with other years, and there are also There has been no major problem with
high rates among Burundian refugees in the deliveries of food to the camp (Appendix 4(f)).
Democratic Republic of Congo (DRC). One problem often mentioned was a recent
The high prevalence of anaemia has been delivery of beans to the camps. These were a
noted in a number of reports. Analysis donation from the EU, a buffer stock of
invariably links this prevalence with malaria. A 5.000MT which had been stored for more than
CDC study team has been in the area recently, 9 months before being donated to WFP. It was a
initiating a longer-term study of clinical aspects 'one-off donation therefore, but the beans were
of malaria control in the region. Furthermore, distributed as part of the ration for 2-3 months.
as a result of the concern over anaemia, They took several hours to cook, even after
UNICEF commissioned an epidemiological prolonged soaking.
study of iron-deficiency anaemia and other There appears to be little documentation of
micronutrient deficiencies in Mtendeli camp in mortality and malnutrition rates in the early

Burundian refugees in Tanzania

months when the camp was being set up and 2 The field study
agencies were establishing programmes. The
estimated population was much higher than the
population counted at the July registration, so 2.1 Use and acceptability of cereal and
mortality rates up to that time would blended food
presumably have constituted an under-
estimation. A mid-upper arm circumference 2.1.1 Food acquisition
(MUAC) screening was done in late December The general ration/SFP: Food is delivered to
1996 by UNHCR nutritionists. The total the extended delivery points (EDPs) by WFP. In
screened was 6,300; a finding of 1,317 children Kasulu district the EDP is Mtabila camp, serving
with MUAC <12.5 suggested that 20.9% were around 75,000 refugees in Mtabila camp 1,
malnourished. The screening team formed the extension, and Muyovosi camp (4km distant).
impression that this was an under-estimate, From there UNHCR takes responsibility for
because some children were not measured, so transport, delivery and distribution at the final
their actual estimate was 20%-25%. The same delivery points (FDPs), implementing through
team noted a very high prevalence of anaemia. partner agencies. Storage at the EDPs in
A nutrition survey (WFH) in April produced a Kigoma is in the form of Rubbhalls.4 In Mtabila
prevalence rate of 20.9% acute malnutrition, EDP there are 11 Rubbhalls, capacity 3,100MT
including a number of children with (allowing for a buffer stock of two months).
kwashiorkor, but the diagnostic accuracy and The general-ration food basket being
overall result have been questioned. Malaria
distributed at the time of the study was as
mortality rates were very high at this time.
In September (at the same time as this study), Cereal grain/or meal:
another survey was conducted, and the result in 400gm/ppd or 350gm/ppd
Muyovosi was a malnutrition prevalence of Pulses: 120gm/ppd
3.9% (WFH <80%). This apparent steep decline CSB: 30gm/ppd
in malnutrition since April suggests that, while Oil: 20gm/ppd
there has almost certainly been an Salt: 5gm/ppd
improvement, the April survey was an over-
estimate of prevalence. The supplementary This is calculated to provide a range between
feeding programme (SFP) and therapeutic 1,950 and 2,100 kilocalories/ppd, depending
feeding programme (TFP) were being handed mostly on whether maize grain or meal is
over from one agency to another in the early provided. The proportionately high ration of
months of 1997 and took some time to be pulses reflects the importance of beans to this
organised, so data from that time are not population, but sometimes lentils are given in
consistent. place of beans.
The monthly health report for August 1997 When both grain and meal are available in
(TRCS Kasulu) gives an average crude the EDP, the meal is given as a priority to new
mortality rate (CMR) of 0.7/10,000/day and arrivals and to the newest camps. Maize grain is
<5MR of 1.8/10,000/day. Mention is made of a given to the old caseload. In Mtabila extension
decrease in mortality due to malaria, and this is near Muyovosi, where the most recent arrivals
explained in terms of improved case live, maize meal is always given, because there is
management. The daily morbidity tally sheet no milling machine, these refugees on the whole
includes anaemia, but no other micronutrient have less access to cash than the more
deficiency. These should be recorded under the established refugees, and the meal is quicker to
category 'other', if and when they present. In prepare and cook. Muyovosi has thus received
interview one of the clinical officers said that he both grain and meal in recent months,
saw vitamin A deficiency only rarely, an depending on availability and priorities. When
estimated three times during his eight months meal is given, it is either white meal, which has
in the camp, but that he had seen many cases of been milled in Kigoma or Isaka from locally
angular stomatitis, mainly in children. He had grown maize, or imported yellow 'cornmeal',
not seen vitamin C deficiency, but anaemia was which is soy-fortified cornmeal (fortified with
highly prevalent. vitamin A and calcium) or unfortified plain
'cornmeal'. The distinguishing marks on the
packaging are a red spot and blue spot
(respectively). The imported meal is packaged

Acceptability and use of cereal-based foods in refugee camps

in 25kg multi-layered paper sacks with heat- some refugees and the Tanzanian workers
sealed inner bags made of polythene, with a doing the scooping. This cheating resulted in a
shelf-life up to one year; while the white meal shortage of some items before the distribution
and yellow meal milled in-country is bagged in was completed, which in turn led to security
jute or nylon sacks, which reduces the shelf life problems. The system did not involve refugees,
considerably. The corn-soya blend (CSB) given except for requiring them to stand in line to
this year has been mostly from USA, but also receive relief supplies as beneficiaries.
included a consignment from House of Manji, Since 1996, the family-group system has been
Kenya. The refugees made a distinction introduced and is much favoured. Now the
between them, complaining about the poor distribution time is measured in hours rather
quality of the House of Manji product. The than days, it is cheaper, and there is much more
vegetable oil is refined and imported. The salt is refugee involvement. The family groups consist
iodised, and procured locally in Tanzania from of families of equal sizes, with a maximum of
the Nyanza salt mine, Uvinza, in Kigoma around 100 people in each group. For each
Region. group there is a list, and each distribution chute
A buffer stock is maintained in the camps of is for a specific family size. There were some
l,000-2,000MT for distribution to new arrivals. problems with this system before registration in
For the last few months, new arrivals in poor July, but since then the lists of registered
condition have been given 50gm (instead of refugees are more up to date and are in the
30gm) blended food/person/day for up to two process of being computerised.
months. The extra CSB was given because of The group leader holds the ration cards
concern at their poor condition on arrival. (plasticised UNHCR cards with serial numbers
Those awaiting verification and registration as printed on and numbered spaces for clipping).
genuine new arrivals and not 'recyclers' are On distribution day, he waits with his group for
given high-protein biscuit (HPB) for a few days. his turn to present all the cards to the Red Cross
The MHA, UNHCR, and implementing supervisors, who check the cards, then clip
agencies are responsible for this categorisation them appropriately. The clipped cards are
and for requesting food from WFP as taken to the designated chute, where a Red
appropriate. Cross supervisor checks the cards against his list
During the early months of 1997, transport of leaders and group size, and then authorises
was problematic because of the state of the the required amount to be weighed out; the
roads, and so distribution was weekly until group leader then signs against that amount.
March, since when it has been every two weeks. The weighers at the chute then weigh each
Two days before the date of the distribution, the commodity and hand the food over to the
Food Committee meets. WFP presents its leader. He calls his group, who help to carry it to
current stocks at the EDP, UNHCR presents the the nearby shelters (still within the well-fenced
official population figures to be served, and so distribution compound), and it is then shared
the total amount to be distributed is calculated. out between the group.
UNHCR then formally requests that amount. The group leaders are selected by the group
Zone leaders and implementing agencies are in a simpler version of the election of zonal
also present at this meeting. The leaders are leaders (1.2). The group members may come
then charged with publicising the ration and from different zones, but they should be
commodities to be distributed and the date of adjacent zones, and the maximum number per
delivery. This includes written notices around group is 100. When asked how group leaders
the camp. are chosen for the distribution of food, some
System of distribution: Up to 1996, the refugees said they choose the one who puts
individual-family system was used in Kasulu. himself forward, who can read and write, and
Each family would collect its ration from the who is active. The agencies tried to get a mix of
distributors at the chute. According to agency men and women as group leaders; but, as with
key informants, there were many problems with zonal leaders, there is a predominance of men
the system. It was very time-consuming, so that (with only about five women group leaders).
for a camp of 13,000 (Mtabila for example) the The philosophy of the family-group system is
distribution took 2-3 days. It was costly, because that refugees take the major part in it. This is
it was labour-intensive, and there was consider- one reason why standard scoops have not been
able over-scooping as a result of deals between imposed in this system, as the scooping is the

Burundian refugees in Tanzania

business of the group, to carry out in whichever distribution provides a place to focus grievances
way they want to do it; the process is observed by over, for example, land or compensation, i.e.
all members of the group. issues originating in Burundi.
Food-basket monitoring should in theory be The SFP beneficiaries are children
conducted every two weeks by a medical <80%WFH and those with oedema. Only
implementing agency, not by the agency pregnant and lactating women who are
responsible for food distribution. This is a referred by the maternal and child health
problem in Muyovosi, where one agency is service (MCH) are included in the programme.
responsible for both medical and nutrition The ration is a weekly dry ration of 1.5kg pre-
programmes, and for the food distribution. mix. The pre-mix is constituted according to a
Therefore an agency from the neighbouring 10:1:1 ratio of CSB:oil:sugar. Vitamin A
Mtabila camp was asked to do the monitoring, supplementation is part of the programme.
but this has not been fully effected. According to The TFP is wet feeding only, using a
a protocol developed by UNHCR, on combination of high-energy milk, porridge
distribution days a systematic sample of around made from pre-mix, biscuits, ubugali, and beans.
40 beneficiaries should be taken as the refugees The CSB used in the selective feeding
leave the distribution area. Then each programme has been of three types, namely
commodity of the selected refugee should be CSB from USA, CSB from House of Manji (both
weighed, and the weight and the ration-card from WFP), and UNIMIX from House of
family size recorded. Later a report is printed Manji, donated to UNHCR by UNICEF.
which gives the kilocalories received by each According to the SFP supervisor, they instruct
person in the sample, and the average amount the women to add the pre-mix to water, ratio
overall. The distribution in Muyovosi takes 1:2, and to boil it for 10-15 minutes.
about one and a half days to complete. During In interview, whenever people were asked
the first distribution during the course of this what they had received from the distribution in
study, there seemed to be some confusion about terms of maize grain or meal, the amount
which agency was responsible for the quoted (in bowls) was almost always less than
monitoring. The reports issuing from both their entitlement. A bowl of grain (weighed by
distributions which took place during the study the researcher) was between 1.5 and 1.8kg, but
are not reliable and indicate that the food- the refugees measure not by kilogramme
basket monitoring teams need training in weight but in tin bowls, which were issued as
analysis of the data and in the standard protocol part of the kitchen sets and are commonly used
to follow. It was also not clear how findings were in the market and as scoops at the general
to be followed up, if and when they showed that distribution. It is not clear why in interviews the
a proportion of families had received less than reported amount was consistently less than the
90% of their entitlement. Even if conducted entitlement, apart from the common tendency
correctly, the exercise only checks that the to under-report assets to someone who is
requisite amount of food allocated, for that perceived to be in a position to influence an
population has been distributed equitably at the increase or decrease in ration. No one in
point of distribution. It gives no insight into how interview volunteered or, when asked, replied
food is used from that point on. that they were unhappy with the system. The
Empty sacks (jute and nylon) have to be common complaint was that the food was not
returned to WFP, which distributes them on enough, but that you could not blame the group
request to the water, sanitation, and community leaders for that. One woman said that it was fair,
services, and for rebagging at the EDP and in that the leader collected it, and they all
Kigoma port. helped to carry it to a place where they divided
Complaints are dealt with on a case-by-case it between them equally in front of everyone
basis, the TRCS and WFP field officers being until it was finished. One interviewee's
present during the distribution. One agency husband, who was currently a group leader,
key informant said that there were inevitably said that of course people would not complain
intra-group quarrels, leading to security about the group leader, because 'you cannot
incidents sometimes, but that the quarrel was argue with the person who will bury you' — i.e.
often over an issue unrelated to food; it was with the person who is in a position to assist you.
raised at the time of distribution because there The same woman who defended the group
would be a big crowd present. In this way the leaders went on to say that there must be selling

Acceptabilily and use of cereal-based foods in refugee camps

on to the market from the store, because the day sorghum for the making of alcohol. The
on which the lorries brought the food was the sorghum was bought in the market from local
day you saw much food in the market. However, villagers. The selling and buying of maize grain
this allegation was not repeated in other were not mutually exclusive activities. One
interviews. She said that the amounts given at woman explained how she had sold white grain
the distribution were not always the same from for 130 shillings/bowl and bought yellow meal,
one distribution to the next, as the bowls were and sorghum for brewing. With the money
sometimes heaped and sometimes not, and she made from selling the alcohol (according to her,
assumed this was because of incorrect weighing at little to no profit), she had bought yellow
at the chute. The quantities of old beans and grain at 100/bowl and would mill it when the
lentils were the same, but the good beans were yellow meal was finished.
not heaped like the old ones. These allegations If maize meal was bought, then it was always
were not repeated in other interviews. the yellow cornmeal; it was more available in
However, the group leader quoted above, when Mtabila extension market, where it is supplied
asked why there appeared to be discrepancies in consistently, while in Muyovosi grain is often
amounts received, said that some leaders did supplied instead (Appendix 4(f)). A number of
not know how to read the weighing scale at the women pointed out that, while they would
chute, so it was possible they might receive less prefer to buy cassava, sweet potatoes or bananas
than they should, while those who could read rather than maize meal, they could get more
and calculate could argue for more if necessary. family meals out of the maize meal than out of
When asked about standard scoops, this leader cassava and bananas for almost the same
said he would like standard scoops for each amount of money, so they bought maize meal.
commodity for each of the family sizes, which Nevertheless, women were buying cassava
would be used at the chute. flour, and it was much in evidence in the
There seemed to be no fixed pattern within market. The white meal which had been locally
households to determine who would collect the produced and distributed in earlier months was
ration. It seemed to depend on availability on disliked (3.1.4); according to one interviewee,
the day, and could be the woman or man or when available it was selling for 5-15
older children, or any combination of these. shillings/bowl, compared with 90-100 shillings
Most women interviewed had small children for the yellow. The WFP market-survey reports
and most had not received any supplementary also indicate a price differential, but not to that
food at the 'hospital', which reflects the selective extent.
admission criteria to the SFP described above. Lentils were bought also by some women,
Those who had received supplementary rations because, although beans 'were far preferable,
said that they kept this CSB separate from that lentils were so much cheaper than beans. Local
received in the general distribution. beans sold in the camps were around 350
Food purchase: The preferred staples of the shillings/bowl, but in the winter season women
Burundians are beans, plantains, cassava and said the • price could increase to around
sweet potatoes, not maize. In terms of the maize 500/bowl, in contrast to lentils at around
grain they receive in the ration, they 50/bowl.
overwhelmingly prefer the white local maize Most people said that they did not buy CSB,
grain to the yellow imported variety; because they did not have the money and also,
conversely, in terms of meal, they prefer the some said, because most people kept their
yellow imported meal to the locally milled, ration and ate it. Some were buying, however,
mostly white, meal. If given a choice of white because the children liked it. They quoted the
grain and yellow meal, some opt for one, some cost as 70-80 shillings/bowl. On the market the
for the other, many favouring a combination price was reported to be 100 shillings/bowl.
(2.1.4). Interviewees did not mention groundnuts,
The most commonly purchased food items but they were being sold in small amounts
cited by interviewees were root vegetables around the camp. They are grown in local
(sweet potatoes, cassava, mahole) and green villages.
vegetables. What they would like to buy, if they Most people denied having money to buy
had money, was predominantly plantains, sugar or rice. One family, where the husband
cassava flour, beans, small fishes, and sugar. was a group leader, admitted that they bought
The only purchased grain apart from maize was rice at 500 shillings/bowl for a child who did not

Burundian refugees in Tanzania

like the other food, and also they were buying quantities supplied to the camps after the
powdered milk for the wife, who had abdominal registration in July. Another is that Tanzania
problems. No other interviewee mentioned had been experiencing a severe drought and
milk, and it was not seen widely available in the food shortages in some areas in 1997, and so
market. demand for grain was high.
Most interviewees seemed to be buying on a The food most sold to get cash for other
small scale only, as and when money was purchases is the white maize grain. The main
available through sale of maize or, much less reason for selling was to get food which they
often, through earnings. In one house, larger prefer, and also to afford to mill the grain. At the
amounts of grain and CSB were in evidence. time of the study, white maize grain was selling
The woman explained that they had bought 30 for 100 shillings/bowl, yellow maize Hour for
bowls of grain (100 shillings/bowl) and 20 bowls 120 shillings. Some were selling at the
of CSB (50 shillings/bowl) at the distribution site distribution site or soon after distribution to the
the day before from individuals selling their market traders. Others were selling from their
ration, not in bulk from the market. She said homes to traders who come to the houses to buy.
that there was a small group co-operating on The WFP market-price survey over the
building a house, pooling food and cooking preceding months shows a fluctuation in price
communally, but it seemed that this was of maize grain. The price rise may have been
probably, at least in part, a small business related to an overall reduction in relief-food
venture. availability following the registration in July,
Non-food items were hardly mentioned as but also may be related to the concurrent
purchases, and in the feedback session the worsening drought and food shortages in
women denied they had money for such things. central Tanzania.
However, the market had a few stalls selling Yellow meal was also sold, but to a lesser
kangas (women's clothes), combs, batteries, and extent than the grain. When asked why, women
sundry items. said that the meal was for eating, but some of the
The shopping was mostly done by women, grain had to be sold to pay for milling.5
but control within the household of any The lentils were also sold and were available
available cash was difficult to ascertain in in the market in large amounts, but the resale
interview. From other informants, from the value was much lower than maize, as there is
reticence of women interviewees to answer the little demand for lentils among refugees and
question, and from agreement in the feedback locals alike; they prefer beans and are not
session, it seemed that control rested with the familiar with lentils. Therefore some said it was
man. One agency informant said that the reason not worth selling them, as they could get very
why the nutritional condition of men tended to little in return and the family would be hungry.
be better than that of women and children was One woman displayed a pot full of lentils, saying
that men could earn cash in local villages and that no one in the family liked them, so she
then they could buy and eat cooked meat and saved them over the weeks until she had enough
other varied foods in the small restaurants in to sell to get a sum of money with which she
the camp or locally. This was not discussed in could buy other things. During the course of the
interview, when men were often nearby or in interviews it seemed that, faced with the choice
the vicinity. between unfamiliar lentils and old beans (which
Food exchange/sale/reciprocal arrangements: took the best part of a day to be cooked), some
The local Tanzanian population grow white families chose lentils, others the beans. In one
maize, they eat it and like it, preferring the household the researcher witnessed a straight
white grain to the yellow, as do the refugees. exchange between neighbours at the rate of 1
There is no market for the yellow grain, while bowl: 1 bowl (which matched the ratio of prices
there is a thriving market for the white grain. of these items on the market).
There is also a market for the yellow meal The vegetable oil supplied is exchanged in
among the camp residents, but local Tanzanians large quantities for palm oil (mabese). There is
are not used to this and so are less likely to buy. much evidence of this at the time and site of
Market-price surveillance from the distribution, and also in the vicinity of the
beginning of 1997 shows a rise in prices of relief market. Interviewees said the exchange rate
grain from July onwards (Appendix 4(d)). One was 1:2 palm oil:vegetable oil. Women in the
explanation for this is the reduction in food feedback session asked for more vegetable oil to

Acceptability and use of cereal-based foods in refugee camps

be supplied, so they could exchange it. Palm oil each; in another house a tailor was at work. One
is used by some for rubbing into the skin and woman was making cooked food for sale in the
hair. market: buswage, a soft ubugali made from boiled
Some agency key informants said that CSB and pounded cassava. Each portion, wrapped
was not sold, except on an individual basis by in banana leaves, is sold for 50 shillings, and is
families without small children, and that it was eaten as a snack and by people travelling.
not seen in the market. Another said that the Locally brewed alcohol made from sorghum
sale fluctuated, so that it was seen sometimes with or without some maize added (indimano) is
and other times not. He said it was not seen in made and sold mostly by women, sitting under a
the early days of the camps but had been seen tree or in the house, at around 25 shillings/small
more recently, and he speculated that maybe cup; or, according to another interviewee, one
local people were developing an interest in it; 20-litre bucket can yield a total of 800 shillings.
but there was no evidence of this. In the feedback session, the women said that
In interviews women denied selling CSB, men did make it sometimes, but they were not
saying that it was too small a ration to sell or to skilled at it.
exchange, and that they all liked it, the children Gifts and reciprocal arrangements were not
especially. One woman said, 'When you get it, mentioned, but one woman remarked that
you keep it; it does not go out of the house', and there were rich and poor in the camp, they were
this sentiment seemed to be shared by most not all of an equal status; and another woman,
women interviewed. However, in the market in when asked about gifts, said that 'the rich ones
Muyovosi it was seen selling at 100 shillings a may give a small amount', but she was not more
bowl (before and after distribution). The specific. Two women mentioned having been
amounts were small compared with the given a small amount of CSB by a neighbour.
amounts of maize flour and lentils, and varied The occupants of the zones are all mixed, and
during the time of the study. On the day of the neighbourhoods in Burundi did not leave and
researcher's visit to Mtabila extension, only one arrive together, so there is probably not an
small heap was seen on sale. This irregularity is overall consistent pattern to inter-household
reflected in the WFP market-price survey of support.
February to June 1997 (Appendix 4(e)). When
Food production: Given the size of the plots in
asked in interview how they thought CSB came Muyovosi (1.1), opportunities for cultivation
to be in the market, women were not are far more limited than in the neighbouring
forthcoming with ideas, but in the feedback Mtabila settlement. Some said that anyway the
session they agreed with the suggestion that it soil was poor and they did not have confidence
was people like bachelors who would be selling; in it. The crops most commonly planted near
and agency staff pointed out that there were still the houses were cassava, lengalenga, ntole,
people with more than one ration card, and onions, tomatoes, pumpkins, and in some cases
there were some 'recyclers' who might have a few beans. A few had one or two banana trees.
access to more than one ration. Some had maize which they could roast when
Apart from work with the agencies in the ripe. In the past they had received seeds from
camps, mostly on a casual basis doing labouring agencies and from the more settled refugees in
jobs, the main means to earn money seemed to Mtabila. Some complained that rats and pests
be trading and bartering with the local villagers, were eating the crop.
acting as agents for local businesses involved in At the time of the study it was very dry; small
the camps, selling alcohol, and a few artisan rains were expected in the coming weeks. Given
activities. It was men who were mostly engaged the season, the plot sizes, and the fact that the
in these occupations, except that production refugees have been in the camp less than a year,
and sale of alcohol was predominantly the work production was not very significant, apart from
of women. For contact with the local farmers the cultivation of green vegetables and some
and traders, a bicycle is essential (and women beans.
were rarely seen riding bicycles). Some artisans
are hindered by lack of tools: a carpenter's wife 2.1.2 Food processing
said he could do only limited work, because he Storage: Food is stored inside the huts; there is
had to leave most of his tools behind. At one no separate storage area. Grain is stored in
house, a man sat outside making large grass sacks, and so is much of the meal and CSB. Jute
mats which were being sold for 800 shillings and nylon sacks are returned to WFP, but some

Burundian refugees in Tanzania

have been retained or bought by the refugees weights of the cereal, test weighings were done
(150 shillings each) or brought with them from for the study, with the following results:
Burundi or Zaire. The maize meal is often 1 bowl of maize grain, heaped 1.8kg
stored in the strong paper sack with polythene 1 bowl of maize meal, heaped 1.4kg
lining which is the packaging of the USA 1 bowl of maize meal, flat 1 .Okg
product and which is shared between the 1 bowl of CSB, heaped 1.2kg
groups. In these sacks, the meal can be kept for
1 bowl of CSB, flat 0.8kg
several weeks.
Only two women were storing the CSB in a Preparation: Food preparation, including the
metal container, one in a bucket, the other in a preparation of alcohol, and cooking is the job of
saucepan. Often the CSB, being a relatively the woman in the house, the wife of the head of
small quantity, is kept in a plastic bag hung from household and/or an older daughter.
the ceiling, or wrapped in a cloth and put in a White maize grain is either milled (2.2.2) to
cardboard box with something heavy on top. flour at the milling machine or pounded and
Cassava can be stored for varying periods, then boiled (2.1.2). The grain to be milled is
depending on the processing (1.1.2). Fresh cleaned and washed first; several women, but
cassava, unpeeled, can be kept for only a day or not all, pound it first before taking to the mill.
two. Dried cassava can be kept in a sack for The grain which is not milled is first cleaned of
about three months, and cassava flour if dried debris by winnowing with a tray, then soaked
well can be kept for up to three weeks. Dried fish for a while in water to make it clean and soft. It
can be stored for several months. is then pounded, then washed again before
cooking. Women said that you could boil whole
In the feedback session, it was said that palm
maize without first pounding it, and this would
oil could not be stored as long as the vegetable
make the maize last longer, but it was then more
oil because it was less refined, and that the
difficult to digest.
longest period to store even the refined
vegetable oil, in the huts, would be about one Pounding by hand is done in an insekulo made
month. of wood, available to buy at around 700
Rats were said to be a major problem. In shillings. Those who did not have one said they
Burundi people had used a poisoned bait. An were borrowing from neighbours when
agency informant said that poison had been necessary. One man showed us the one he had
considered, but it was risky in camp conditions, made for his family from a local tree. He said
and refugees said they did not use it here, that not everyone knew how to make them, so
because it was not available and/or was not some people borrowed theirs and used it for
affordable. They said there were many rats, free. He said they did not charge, because 'you
because the camp was recently a forested area. have to help each other'.
In the feedback session women mentioned ways The insekulo is also used for pounding some
of killing the rats, including pouring boiling green leaves such as cassava leaves, which are
water down their holes and then killing them first roasted, then pounded and fried with
with sticks. onions.
Most women said that they milled at least
Measurement: The standard measure used at some of the white maize grain, but one said they
household level, in the distribution and in the never did, and only made ubugali from the
market, for small amounts, was the tin bowl; for yellow meal they were given. The white maize and oil it was the tin cup — both grain in this case was either pounded and
provided in the kitchen sets distributed earlier cooked whole (as ndete — see below) or sold to
to the refugees. For making ubugali and get yellow maize meal or cassava flour. The
porridge, women said they knew how much yellow grain was milled, but not pounded to
meal to add according to how much they make ndete. Women in interview said that it did
needed to make. One example given was for the not pound well and broke into pieces; this was
cooking ofubugali for two adults and a child: the confirmed by women in the feedback session.
woman had used one cup plus a handful of There was no evidence of grinding to flour
meal. In the cooking demonstration (CSB), with stones at household level in the camps. One
both women measured out the CSB with woman said they used to do this sometimes in
dessertspoons. Burundi if they had not enough money to go to
The bowls at distribution are sometimes the mill, but she gave the impression this was
heaped, sometimes flat, so, in order to clarify not a routine activity. She said there were no

Acceptability and use of cereal-based foods in refugee camps

suitable stones for milling in the area around Two processes to make it into flour were
the camp. In the feedback session, women said described. One was to peel the fresh cassava
that they had stones in Burundi and some did roots, then cover with banana leaves for two
use them regularly when money was short. days to ferment; this cassava then looks black or
They said it was the young ones who did not grey on the outside. It is then dried, cut into
know how to use them; the old ones knew, but small pieces, and milled. The other way was to
there were no suitable stones in the vicinity. peel, then soak the cassava root in water for
The maize meal requires no preparation and about three days, then dry and cut it into small
is cooked as ubugali, although one or two pieces, and then mill it. Another method of
women mentioned sievingfirst.A small number preparation was to peel and then soak the
of women interviewed also said that they sieved cassava over several days, changing the water
the CSB. One said she had sieved only the from time to time. Then, when the cassava is
coarse one which they had received several sweet, it is boiled and then pounded to make
months previously, but others said they also buswage.
sieved the finer one which was being distributed The small dried fish (indagara) which are so
currently — but not the one given at the popular locally and among the refugees take
hospital. These women said that they threw almost no preparation and are cooked quickly
away the remains in the sieve, but in one in a little water and oil.
household they were observed to be pounding To make indiviano from sorghum (it is mostly
what had not gone through the sieve, intending made from sorghum, but also maize grain may
to make porridge from these pounded remains. be added), the grain is soaked in water for a day,
In the feedback session, women said that many put in a container for three days to germinate,
had sieved the old CSB and some the good one then dried in the sun and taken to the mill. The
to remove the dirt. flour is then mixed with hot water and left for a
Women, when asked, described the process day, when it is ready to drink. At this point it is
of preparing porridge when they were at home said to be not alcoholic: it is 'like a juice', but it
in Burundi. They used to make it from sorghum may be mildly alcoholic. To make it alcoholic,
or millet. This had to be pounded, dried, and fermented flour is added, and it is left for
then ground to a flour. They might also make it another day or so. One woman said that she
from cassava flour, or a mixture of this and would like to make this juice from bananas, but
grains (sorghum and millet) and sometimes that these were expensive and the process
soya. However, women said that they did not required more equipment, which she did not
make porridge from maize meal alone. have. The banana juice' is made into a strong
The beans were cleaned by hand, checking liquor called ugwagwa, when fermented
for stones and other debris, and then washed sorghum is added.
prior to soaking. This process was longer with Cooking: The white maize grain, but not the
the old beans, because they were also checking yellow, was used to make ndete. This is pounded
for spoiled beans among them. The preparation maize, boiled and mixed with beans or lentils. If
of the old beans would begin the evening the old beans were being used, they were put on
before: they were soaked overnight, then put on the fire in advance of the maize grain, but if
the fire in the morning to be eaten later that day. lentils or the good beans were to be mixed, the
The lentils were cleaned and washed before grain would be put on the fire first and then the
cooking, but not left to soak. Several women beans or lentils added. The grain would take
were also kneading the lentils in water to detach two-three hours to cook, so the beans or lentils
the outer skin, then drying and winnowing to would be added after about two hours. Before
remove the skins. This, they said, was done to serving, salt and oil would be added.
improve palatability and digestibility. Apart from deep-fried snacks made from
Cassava is a very popular food among the maize meal which were selling on the market,
refugees and is locally available and seen in all maize meal in the camp was used exclusively for
the markets. It was being prepared in a number ubugali, a thick stiff porridge which can be made
of ways, depending partly on type (whether it is from cassava flour also. Ubugali is distinct from
bitter or sweet cassava, as the bitter has to be umusululu, which is a thin porridge. Water is
soaked). The most simple way was peeling the brought to the boil, then flour is added, and it is
fresh cassava root and boiling it whole like a stirred on a hot fire vigorously and
potato. The other way was to make it into flour. continuously; the consistency is easily

Buriindian refugees in Tanzania

controlled by adding more flour or water as the The cooking method used in the
porridge starts to cook. After about 10 minutes demonstration was simple: adding some flour to
(depending on amount), when all the water is cold water and then bringing to the boil, stirring
absorbed, the thick dough is turned out on to a gently until cooked. Sugar was added if
plate ready for eating with a range of dishes if available, halfway through the cooking process,
available, such as meat, fish, green vegetables, but nothing else. One woman said that in
and beans. Most often in interviews, the ubugali Burundi she would sometimes add an egg also.
was to be eaten with beans or lentils from the In the feedback session, women said that some
ration, and lengalenga. children liked to eat CSB raw, and did not have
When asked if they mixed cassava flour with any subsequent discomfort or diarrhoea. Some
maize meal, a few said they did, at a ratio of 1:3 children were witnessed eating it raw at one of
cassava:maize, but some said they had had the distributions observed.
cassava flour only occasionally since coming to Most said they had not received any cooking
the camp. instructions about CSB, one woman saying it
The old beans or kambaranga which had been was not necessary, as they knew how to cook
given by WFP some weeks back were universally porridge. They were not aware that it was pre-
acknowledged to take an extremely long time to cooked. The consistency of CSB porridge was
cook; it was believed that insufficient cooking medium thickness (one woman said about 1:3
would cause abdominal problems. A few women flounwater), so that it was eaten from a cup with
told us that they put a small amount of soda a spoon, but was a little runny. Sugar was added
bicarbonate in the beans and that this shortened if available, and was certainly preferred, but
the cooking time to one or two hours (i.e. to nothing else was added. However, in the
about the same time as for normal beans). One cooking demonstration, the proportions were
said it did not change the taste, but another said about 1:5, and the resulting porridge was quite
the taste of it was the reason why she did not use runny. Women said they preferred the taste of a
it. Some had never heard of doing this. One thicker porridge, but made it more runny to
woman said she did not use soda bicarbonate, conserve the CSB.
but added ash as the beans were cooking to Oil is used not in ubugali but in the dishes
soften them. which are eaten with ubugali, such as the beans
The lentils were boiled and their and green vegetables. Some reports (WFP
acceptability, especially to children, was 1995) indicate that the diet of the Burundians in
improved by adding green vegetables with oil. Burundi is normally very low in fat, partly
The cooking time was much shorter than that because they produce so little oil. However,
for the old beans, and shorter also than the when this was discussed in the feedback session,
cooking time of the good beans. Salt and oil women said they liked and used palm oil
were added to the lentils before cooking was normally. They had palm trees in the lowland
completed. areas and made their own unrefined oil. People
It was not easy to establish in interview the in the highlands did not have this access, and
length of the cooking time for porridge tended to use cow fat instead. An agency key
(umus'tilulu). Women were unable to tell us in informant who had eaten food often in the
minutes, but said that it was the same cooking camps remarked that the use of oil was evident
time, or even more, for CSB porridge as it was in their cooking.
for the sorghum porridge they used to cook, Groundnuts, having been roasted and
one saying that millet porridge was also faster pounded, were also added to vegetable dishes in
than CSB. One or two said it was about the same Burundi. They were for sale in the camps in
time as for cookingubugali (around 10 minutes). small amounts, so are used by some in cooking
Observation in the camp suggested a shorter and/or eaten as a snack. However, groundnut
time of 5-10 minutes, and a timed cooking oil was not known, palm oil being the customary
demonstration revealed a cooking time of five and preferred oil for cooking.
minutes for an amount of about three portions. The fuel-efficient or 'improved' stoves (2.1.5)
In the feedback session and demonstration, were very much in evidence around the camp
women said the time depended on the amount and in the houses visited. When asked, women
of firewood used, i.e. the strength of the fire, confirmed that the stoves consumed much less
and repeated that they cooked CSB for the same firewood than the three-brick fire and gave
time as they would cook any porridge. examples. One woman said that for the old

Acceptability and use of cereal-based foods in refugee camps

beans, which were cooking during the afford to mill, and/or milling facilities were not
interview, she was using only about 20 pieces, yet established. Some said it gave them
whereas before it would have taken a large discomfort and it was not something they were
bundle. She said the stove was also better used to eating in Burundi. Some said the young
because food could be left cooking on the fire children ate it but the older ones refused,
while other jobs were done, as it did not need to because they were used to cassava, bananas, etc.
be tended, unlike the three-brick fire, and it was The women in interviews seemed to be
safer from the wind. The stove was new to them implying that it was an inferior food, and, in
when they came to the camp, she said, but they some cases, was a sign of hardship. In the
had been taught how to use it and it was very feedback session, women said they tolerated it:
good. Most are portable, so that, when the rains it was not their first choice of food, but they did
begin, they will be able to carry their stoves cook it sometimes in Burundi. There then
inside the hut. Some were cracked and in need followed a discussion of the regions of Burundi
of repair. A number of people had a three-brick where people were more likely to eat maize and
fire inside the house, which they said was mostly particularly ndete, compared with other regions;
for light in the evening and some said for Mugamba region was specifically mentioned as
porridge in the mornings. a region where they did, while Bujumbura was a
Cooking sets had been given in June/July. region where they did not. The reason given
Before that, most people used traditional clay was that in Mugamba they grew a lot of maize
pots or others, if they had the money to buy. On and their cassava did not thrive as in other
arrival, they had been given jerry cans for water areas. This discussion highlighted the existence
storage. of distinctions between regions made by the
refugees in relation to agricultural and food
2.1.3 Intra-household food distribution practices.
Most women interviewed said that they had
Almost all the interviewees said that the
eaten in the morning after waking, and mostly it
whole family liked and ate the CSB porridge in
was CSB porridge which they had prepared. If
there was cooked food remaining from the the mornings; throughout the camp, people,
evening before, then the household members including men, were seen to be eating it.
would eat that instead of or in addition to the Women said that older children asked for sugar
porridge. Sometimes between 9 and 10 in the in it, and all would like to add sugar; but if they
morning the preparation of porridge was had none, they still ate the porridge. By the time
observed, so it was not necessarily taken first that the next distribution came, almost all
thing in the morning. By mid-morning they households had little or no CSB left. In two
would start cooking the old beans, if they were houses, however, they still had quite a lot of
having them for supper with ubugali. Otherwise CSB, even though it was just before distribution
they would be planning to cook ubugali and was due. One woman said it was because at the
lentils or green vegetable for lunch or ndete. moment only the children were eating it, and
This would be eaten some time between noon another that the children were refusing it
and 2 p.m., and would fit to some extent around because there was no sugar to go in it, and that
those children coming in from the morning they did not feel like eating it when they were ill.
shift at school. The evening meal after dark In this household the mother was ill on the day
would be ubugali or more of the ndete. While of interview, the children thin and scabied, and
ubugali, if eaten twice in a day, would usually be there was sorghum drying outside in
cooked afresh each time, ndete and beans might preparation for making alcohol. The only other
be cooked once and then eaten once or twice on people declining CSB were occasionally
that day and even the following day. Apart from pregnant women, who said it gave them
the ration food, fresh cassava and sweet potatoes indigestion.
were also being bought and eaten to vary the
Although women tended to say in reply to
daily diet. Most women did not volunteer this
questioning that the children ate what the
information, but preparation and cooking of
adults ate and there was nothing in between,
these foods were observed around the camp,
small children around the houses were
and they were always available in the market.
observed to be eating sweet potatoes, sometimes
Ndete was eaten by some but not others. Some as a snack, 'something while they are waiting for
said they had cooked it when they first came to the meal', or finishing the remains of an earlier
the camp, presumably when they could not meal. Women also said that if the children asked

Burundian refugees in Tanzania

for porridge a second time during the day, they are three 'plates', one for the men and older
would prepare it for them, and some did boys, one for the woman and children, and one
prepare it routinely for the late afternoon. just for the very small children. According to the
Others said there was not enough in the ration same informant, traditionally there were foods
to cook it more often than once in the day; reserved for men alone and some avoided by
others said that the supply was usually finished women, but nowadays these practices are no
within a week. longer observed.
The drinking of tea was not mentioned In interviews in the huts, women said that in
spontaneously in interviews, but, when asked, the morning and evening family members ate at
some said that they used to take tea in the early the same time, including the woman who had
morning in Burundi, but not here, because they cooked. The woman allocated the amounts
could not afford to buy sugar. However, it given and then would usually eat from the same
seemed that the taking of indimano was more plate as the children, the men and older boys
common than drinking tea, even in many of the eating from a separate plate and in a different
households when in Burundi and now similarly space. In one relatively wealthy-looking hut, in
in the camp. Those interviewed said that its own grassed compound, the woman said that
indimano was taken by all members of the family, she and her husband shared a plate, the older
including children, that it was like a food and children (boys and girls) shared, and the ones
not alcoholic. There was no special time of the below 9 years shared a third. One woman
day to drink it; it could be taken at any time. explained how she now had her own mother
Oil was eaten by everyone; according to the staying with her and so could not eat with her
interviewees, even young children would be husband. The older boys, her husband, and the
given oily food. There were no food avoidances grandfather sit together, she sits with the
or taboos for pregnant women reported in children. In the feedback session women
interviews or the feedback session, though one explained that, where small children had a
young refugee informant working for one of the separate plate, the mother would be close by to
agencies said that the older generation used to ensure that the children were eating and to add
say that pregnant women should not eat any more if needed.
part of a chicken, and generally all the best parts
of the animal were saved for the men in the 2.1.4 Food preferences
family. While this did not arise in interviews, As already noted, the preferred staples of the
where meat was never mentioned as an option, Burundians are beans, plantains, cassava and
one woman described how her husband would sweet potatoes, not maize. If they have to
divide up the food after distribution, selecting a receive maize grain, then they overwhelmingly
portion of it to be sold to get alternative foods prefer the white (local) maize grain to the yellow
for him and a child, as they did not like the (imported) maize grain. According to the WFP
ration food. The remaining portion was for Field Officer, if yellow and white were both
consumption by her and the other children. given at a distribution, each had to be shared
This meant that at times she had to cook two equally among the refugees, so usually they
types oiubugali. In the feedback session, women gave first one for seven days, then the other for
said that they did not give the best to the men, seven days. He said that they always gave the
because since they had little, they had to give yellow grain first with this split system, because,
equally. if they gave the white first, the refugees would
not return to collect the yellow. Some women in
Breastfeeding continues up to 2 or 3 years, interview said that the yellow grain could not be
and weaning begins at around 5 months. pounded to make ndete because it disintegrated,
Commensality: According to a young female and that, when milled, the taste is different from
Burundian refugee working in the MCH clinic, and inferior to both the white grain when
there is a difference between the 'educated' and milled, and the yellow meal (pre-packed). Local
the 'non-educated' families. In the latter, the Tanzanians do not like the yellow grain, they
men and older boys are given a big share by the are not used to it, and so there is no market for
woman and eat separately, while the woman it. It constitutes only a small percentage of the
and all other children eat together from the total grain distributed.
same plate; the younger ones lose out, because After the white grain, for which there is a
they cannot eat as fast as the older children. In thriving market within and outside the camp,
the 'educated' families, on the other hand, there the refugees prefer the yellow maize meal

Acceptability and use of cereal-based foods in refugee camps

overwhelmingly over the white maize meal. maize meal. The Tanzanians prefer the grain
This apparently contradictory preference which they can mill themselves.
(considering that they prefer white over yellow While maize was not an everyday staple of
grain) is explained by WFP partly in terms of the these people when in Burundi, many of them
very fine texture of the yellow maize meal, grew some for sale, and a few had made ubugali
imported from the USA, and its superior from maize meal. But their stated preference
packaging (25kg paper and plastic heat-sealed for everyday staples was always the range of
sacks). Some yellow maize grain is milled in cassava, plantains, sweet potatoes and, of
Isaka, but this is a relatively small proportion. course, beans. Their preference for ubugali was
Both cornmeal and soy-fortified cornmeal ex- also cassava flour, but this was more expensive
USA are distributed, the only distinguishing than maize meal, as more was needed to satisfy
features being a blue spot and red spot hunger. The most expensive of these altern-
(respectively) on the outside of the paper sacks. atives was the plantains, and people said they
The WFP monitor in Muyovosi said that they could not afford to buy them. The other highly
were distributed as the same item, but he had desired item was indagara, small dried fishes, of
heard refugees saying they preferred the taste which there were plenty on the market. Some
of the 'blue spot' meal over the 'red spot'. When also mentioned rice, especially those from
asked at interview and at the distribution site lowland areas where rice was a more common
about this preference for the 'blue spot' over the staple. One man who said he bought it regularly
'red spot', some refugees made the distinction, said it was because they used to live near the lake
others did not. In any case, the number of'blue in Burundi where rice was plentiful. However,
spot' sacks is small compared with the 'red spot', most people did not mention it unless
so most were receiving the latter and did not prompted.
complain. In the feedback session, when asked, The CSB distributed in the general ration is
women said that in the past the 'blue' had been almost all from the USA and is well liked by the
better, but nowadays it was all the same. refugees. A small amount of the CSB ex-House
Some of the women interviewed said they of Manji consignment (2.2.3) had recently been
preferred the white grain to the yellow meal, given in Mtabila 1 for seven days, but in
but not all. Most preferred the yellow meal. In Muyovosi it was given only in the SFP. The
the feedback session, women said they liked a texture was noticeably coarser than that of the
combination, presumably because this allowed CSB from USA. Almost all expressed a
more flexibility. The ready market for white preference for adding sugar to the CSB, but the
grain means that it can be quickly converted to absence of sugar did not make it unacceptable.
cash; but particularly for those who depend A group leader stressed how at distribution,
most heavily on the ration, the meal is when dividing up the food, 'everybody wants
preferable, because it requires no milling, and the CSB' — meaning that they wanted it for
the taste and quality seem highly acceptable to consumption, not for trading. One woman said
all. Some added that, unlike the other items in that as soon as they tasted CSB, they liked it
the ration, the yellow meal was 'healthy' and did straight away. This same woman said of the
not make them ill. It was 'sweet', and they ration food generally, 'If you are not used to a
preferred it even though they never had it food, at first it is difficult, but then you get used
before in Burundi. There was an observable to it.'
contrast in the general atmosphere at the The lentils distributed are unfamiliar to the
distribution where grain and lentils were given refugees, who unanimously prefer beans of any
and the distribution where meal and beans were type. Women have tried to make lentils more
given, but this was at least partly due to the acceptable to household members by peeling
receipt of the good-quality beans. before cooking and also adding green
Many refugees said that the white maize meal vegetables and oil to them. As the resale value
which is ground in Kigoma and Isaka was was low, some women said that they cooked
rough, not sweet like the yellow, and some said them on days when there was no alternative
it contained glass and sand (refuted available ('I don't like them, but I eat them'), or
unequivocally by WFP, who however would swap them for the old beans, which took
acknowledged the strength of these rumours); so long to cook. Some said they preferred the
others said that it had a sour taste and bad lentils to the old beans; the preference was
texture. There is very little market for this white related, at least by the women in interview, to

Burundian refugees in Tanzania

the degree of abdominal discomfort which each probably a combination of things, but many
gave, rather than the cooking time. One woman seemed not to know that maize was the main
ate the lentils herself, despite disliking them, ingredient; though they liked it, most did not
giving preferential allocation of beans to her seem to know that it was fortified and partly pre-
children. cooked. Apart from those who had received it at
The vegetable oil was exchanged at a ratio of the hospital, and one or two who had been
2:1 for palm oil, by many people but not all. For shown it at the registration centre on arrival,
some the vegetable oil was acceptable enough. most said they had not received any instructions
However, in the feedback session, women, about how to cook it.
when asked why they preferred palm oil, said it
Fuel-conservation measures: In April 1997, a
was sweet; they were used to it and never got
programme to promote fuel-efficient stoves was
tired of it.
implemented; by August it had achieved almost
2.1.5 Local factors influencing food practices 100% coverage in the camp. At first, people
were given incentives to make and use the
Previous experience of refugee status: Several
people interviewed for the study had been stoves. The circular stoves, with an opening for
refugees in camps in Zaire for up to four years firewood, are made from a combination of clay,
before arriving at Muyovosi in October, soil, ash and grass. They are sculpted over a
November and December 1996. Those who had period of several days, so that the material has
this experience were already familiar with time to dry. It does not need to be fired before
yellow maize flour and CSB and other relief- use. They are portable and were used recently
food items. Their acceptance of the food and in the now-disbanded Rwandan camps, where
their adaptability were presumably more they were given the name cananke, which means
advanced than in the case of those who had 'burning small'. The implementing agency have
come straight from Burundi. However, those calculated that use of these stoves gives a 40%
who had been in Zaire tended to say that they saving on fuel. They mobilised the refugees
had received a better deal there, some saying initially by pointing out the advantages of these
they had been givenfish,among other things, as stoves, which include the facts that they need
part of the food basket. One woman who had less firewood, produce less smoke and no
spent three years in Zaire said that, when they sparks, and cause fewer fire accidents; also the
first came to Muyovosi, they were given a lot of sides of the cooking pots do not get soot-
food, but that now it was less. She asked, 'Is that blackened. As people began to use the stoves,
because they are getting tired of us?' the agency also gave advice about aspects of use;
Those who had had some business in the for example, they advised people to soak beans
camps of Zaire or in Burundi, such as tailoring, before boiling and then to remove them from
selling cooked food, alcohol, or making mats, the fire after some time and pound them, to
tended to be trying to resume these activities in speed up the cooking process, before returning
the camp. them to the fire. They also suggested the use of
soda bicarbonate in the beans to reduce cooking
Understanding of health and nutrition issues, time, but were not sure of the effects of this
including CSB use: Many women said they practice on the nutrient value. They
were used to making porridge in Burundi for encouraged the pounding of maize before
the whole family; but also some spoke of a type cooking it, and the use of lids when cooking;
of porridge made from a mixture of grains and they urged people to refrain from lighting fires
soya bean which had been promoted in the to warm themselves (but to use embers), to put
health clinics there, for those children who were out the fire as soon as cooking was completed,
thin or ill. In the feedback session, women and to have all the necessary food and
confirmed this, saying that it was called mumlaki equipment to hand prior to lighting the fire for
and included wheat meal, milk, and sugar. In cooking.
interviews, most women said they had formerly
made porridge from millet or white sorghum, Markets and local production: The
preferring red sorghum for alcohol. Red surrounding area is agricultural, and the food
sorghum was considered to be nutritious and crops produced and consumed are the same as
'good for the blood'. those in Burundi, and so are familiar to the
When asked what CSB was made from, refugees (1.1). Villagers visit the camp to trade;
women said they were not sure, that it was exchange of food for labour also occurs in

Acceptability and use of cereal-based foods in refugee camps

nearby villages. There is a history of trade are single-piston diesel engines and one is a
between Burundi and this region of Tanzania, twin-piston diesel engine. (In Mtabila there are
so trading and reciprocal arrangements are four single- and one twin-piston engines.) They
already in place. There is very little tension are all commercial enterprises (refugee-owned),
between the local and refugee populations. permitted by MHA to operate within the camps.
There are thus a number of ways that different The machines were bought in Tanzania, and
foods can be obtained to vary the diet, not to the spare parts are available locally. They are of the
extent enjoyed in Burundi, but to some extent. same type as the local village mills.
The capacity of the single-piston engine was
2.2 Opportunities for cereal fortification said to be about 1MT daily (12 hours) and that
of the double-piston engine 3MT daily (12
2.2.1 Household cereal fortification hours). For the four mills in Muyovosi camp,
Household-level milling: It appeared that the this gives a total capacity (assuming a 12-hour
Burundians in Burundi usually took maize and day) of 84MT over a two-week period, or twice
other grain to the mills. Some milled in the that if the mills operated 24-hour days. The
household, using stones, particularly for small amount of grain distributed in Muyovosi camp
amounts of millet, but not routinely for all grain. on 3-4 September 1997 (during the field study)
They had not brought stones with them to the was 157MT.
camp, and the stones are not available locally, so The minimum load required to start the
they were taking most of the grain to mill. The machine is about 3kg for the smaller engine,
only alternative was to pound the grain and boil and 15kg for the larger. The only advantage to
it. Fortification cannot be done with unmilled the refugees of the larger machine over the
grain (including grain that has only been pound- smaller is that it is quicker, but with small loads
ed). Therefore with such practices, household- this is not a significant time difference. The
level fortification at the milling phase would not texture depends on the sieve used.
be an option for this population. The price at the time of the study in all the
mills was 30 shillings per 1.5kg of grain, which is
2.2.2 Cereal fortification in-country
slightly more expensive than the price in the
In 1995/96, before the Rwandan repatriation, surrounding villages, but these villages are not
WFP could not keep up with the demand for near enough to make it worthwhile for refugees
maize meal (Lacy 1997). Since December 1996, to go there to mill. This makes the cost to
US food donations to WFP for the Tanzanian refugees around 20,000 Tsh/MT. It also means
camps have included a larger proportion of that, crudely calculated, about 1.5kg of grain
milled maize than formerly. The advantage is of has to be sold to pay for the milling of a full
course that milling by the refugees is not ration of 5.6kg (14 days per person). Thus the
necessary, but the disadvantage is the shorter amount of meal obtained is 4.1kg, in
shelf-life and the risks of infestation. The paper comparison with 4.9kg obtained when meal is
sacks are damaged in transit more easily than given in the ration.
the jute bags containing grain (Cawkwell 1997).
Locally produced maize grain is procured Regional-level milling: Kigoma town:
within Tanzania, but is issued for distribution Relatively small amounts of locally procured
only inside Tanzania. It is not sent to the other maize grain are milled by WFP in small mills in
WFP programmes in the Great Lakes Region. Kigoma town. At the time of the study, WFP had
contracts with three local mills. Basic details are
Camp-level milling: There are no camp-level presented in Table 3. The cost to WFP at the
mills run by WFP in Kasulu district. Formerly time of the study was 14,000 Tsh/MT.
WFP had an arrangement with TRCS, which Milling times quoted by the operators
operated mills in Kibondo district, but this no depended on the quality of the grain. They also
longer applies. In Nyaragusu camp (Congolese said that the yellow grain milled faster than the
refugees) in Kasulu, only maize meal is now white, partly because it was drier than the white.
given to the refugees, as there is no mill in the However, milling engineers who have visited
vicinity. When grain was given formerly, the the region and produced recommendations
refugees had to leave the camp to reach the mill, regarding operational procedures have
which created some security problems. commented that some grains, including the
There are four mills in Muyovosi camp, and American Yellow, are 'extremely hard and as a
five in Mtabila. Of the four in Muyovosi, three consequence take longer to mill' (Stone 1995).

Burundian refugees in Tanzania

Table 3: Regional-level mills in Kigoma town

Name of mill Number mills Power supply Size of engine Capacity MT/hr/mill Sieve size
Kudaba 2 diesel 14 hp 500-800kg 1.5111111

Kitongoni 1 electric 14 hp 500-800kg 1.5inni
Tuniahini 1 electric 14 hp 500-800kg 1.5iiiin

Other mills have been contracted, and kva), to be operated and managed by WFP
contracts have lapsed, partly because the price Isaka. For a diagram of the containerised grain-
could not be agreed between WFP and the mill milling system and proposed milling set-up, see
owner. At the time of the visit for the purposes of Appendix 4(g). The system could be
this study, Kitongoni mill was not functioning, streamlined more if the grain was delivered in
because there was electricity rationing in town bulk and not already bagged, to save double
and at that time of the day there was no power. handling and bagging.
The mill owner was not allowed by the town The proposed milling set-up could be
authorities to install a generator, because the modified at modest relative cost to include
mill is in a residential area. The Tumahini mill fortification. The vit/min pre-mix could be fed
was temporarily under repair, because the sieve through a dry chemical feeder mounted on the
had been damaged. flour conveyor.
The milled maize from the mills in Kigoma When fully operational, the Isaka milling
was sent mostly to the camps in Kasulu district, facility will be producing 200 MT/24-hour day,
with priority for all meal to the Congolese i.e. maximum capacity will be 6,000 MT/month.
camps of Nyaragusu and Lugufu, where milling The monthly requirement (if maize meal is
facilities are not available. given) currently for the seven camps in Kigoma
region is around 3,192 MT (@ 350gm/ppd for
Isaka (Shinyanga region, Kahama district): 304,000 population). The approximate cost to
Isaka is a major cargo centre for the WFP 'Great WFP, once the mills are fully operational,
Lakes Regional Cluster', which includes should be around 9,000 Tsh/MT.6
Tanzania, Kampala, Rwanda, Burundi, and
The main constraints on milling and
eastern DRC. Food is brought from the port in
fortifying grain for the WFP caseload in the
Dar es Salaam by rail to Isaka and Kigoma.
Great Lakes cluster are not technical or
From Isaka it is trucked to the EDPs in Kigoma
financial, given that WFP has already invested
Region (mostly Kibondo district) and Kagera
heavily in the Isaka milling operation. The USA,
Region, and from Kigoma port (being up-
as the major donor, has recently sent an
graded with assistance from WFP) ferried by
increasingly larger proportion of maize to the
boat to Burundi and DRC, as well as taken by
region as meal, rather than grain. Where this
truck to camps in Kasulu district of Kigoma
was supplied (unfortified with micronutrients),
CSB would need to be continued as the main
Towards the end of 1994, UNHCR and WFP vehicle for micronutrients.
procured mills to be operated in Isaka by the Quality checks are conducted by African
National Milling Corporation (NMC) to mill Marine Surveyors Ltd.
locally purchased grain. The mills were
prototypes, designed to UNHCR specifications, 2.23 Production andfortification of blended food
and with the generators were transported from Most of the blended food used for the general
UK. At the time of this study, the contract ration in the camps has been CSB (donated by
between WFP and NMC had been the USA).
discontinued. The mills had not operated to 2.000MT of CSB ex-House of Manji Nairobi
capacity since being installed, and had sustained was procured by the Regional office of WFPjust
damage due to poor maintenance and lack of before the sudden repatriation of the Rwandan
expertise in operating them. Output had refugees. The blended-food supplies were very
declined to c.lMT/hr, whereas the capacity of low by October 1996, hence this regional
each mill is c.4MT/hour if well maintained. procurement. The October consignment was
Currently WFP is rehabilitating the machines subsequently stored in Mwanza for several
and re-installing two (1 x 240 kva and 1 x 350 months, and the quality is now having to be

Acceptability and use of cereal-based foods in refugee camps

checked, as the contents of many bags were consumed than sold, and many women
reported to be caked. At the time of this study, preferred to be given meal rather than grain.
WFP was undertaking a quality-control However, meal which had been locally milled by
investigation of this consignment. This was the WFP before distribution was far less acceptable
second time that CSB had been procured to the refugees than the yellow imported maize
regionally, the first being earlier in 1996. There meal. These factors are relevant to
had been some problems with quality then also. consideration of a possible strategy of cereal
UNICEF gave 90MT of UNIMIX ex-House fortification in-country either before or after
of Manji to UNHCR early in 1997, but this was distribution to refugees.
used exclusively in the selective feeding
programmes, not in the general ration.
Conclusion 1 'Congolese' refers to residents of the former
Zaire, now Democratic Republic of Congo.
Blended food was widely liked and consumed 2 For purposes of comparison, the standard
by this population and recognised as a plot size in urban Tanzania = 25 x 35 metres,
nutritious food, though they were not aware and in rural Tanzania = 35 x 7 0.
that it was pre-cooked and fortified. Arguably its 3 Hblevel<llg/dl.
appropriate use could be enhanced by 4 Large, pre-fabricated store tents.
dissemination of information on these points. 5 The UNICEF/TFNC survey found only 5%
Maize, the main staple supplied in the ration, of respondents reporting selling part of their
was not unfamiliar to most people, but it was not food ration (household questionnaire).
one of their preferred staples, which were 6 Note approx. costs therefore of milling/MT:
beans, plantains, and cassava. Therefore, when For refugees at camp level 20,000 Tsh/MT
supplied as whole grain, a proportion of it was For WFP at Kigoma level 14,000 Tsh/MT
regularly sold to obtain other food items. When For WFP (own milling set-up) 9,000 Tsh/MT
supplied as meal, it was more likely to be (excluding capital costs of set-up)

Appendix 4(a): Map of Kigoma Region, with location of camps

• /

International boundary

District boundary

Refugee camp S
District capital






40 km



Acceptability and use of cereal-based foods in refugee camps

Appendix 4(b):
Age and sex composition of population, Muyovosi camp



a m
.C E
01 U
e •H

E 0

o £ A

a >•
U rH O
S H O vo in en rH VO vo n o o o fH
4J rt
vo •0
A u
a l-i

3 O
in in •»n r- ci in r»
vo d r- m r- in Cl IN ro
V b.

a >

>, U o PI en n r- rH ci Cl n o o o
o PI H Cl H ^
vo •a
rH fH rH rH

A u
U 3
l*> en i n in n in in a> i n VO CT\ ro d IN

o vo as o\
r~ n Cl
vo PI


[-. m m en o ^
vo r* o
i n vo 0 0

VD o
o in

b. in


, <N o m in o\ ^, 03 rH Cl 00 VD rH n o>
in n fl f^ in 03 Tf vo o
rH in f^ j^» in ^ i d
p) r- n
£ in

in in ^, o m
03 in Cl 00 o 00 rH tH d
v IN f^ M rH fH rH in o o in PI H
H m (N n m CJ rH H d rH d
b. d

v vo O m en f t f^ o <n in vo VD in rH 00
01 Cl vo m i n (N N o o o m n m
** H <*> o m n Cl rH r-t Cl r H d
£ d


i 2i

o u a o s o
CD U b. fH * £

Appendix 4(c): Refugees Statistics Report on Registration for
Kigoma Region, Tanzania. 19-26 July 1997
1. Detailed camp breakdown

Population REDUCTION
(District Camp Nationality After Registration
Before registration In Number Percentage
Kigoma lugufu Congolese 46,065 34.818 11,247 24.41V.

Kasulu Nyarugusu Congolese 47.927 39.211 8.716 18.1
Mtabila Burundese 66.651 44.040 22.611 33.9
Muyovosi Burundese 40.999 27.750 13.249 32.3

SUB-TOTAL (KASULU) [ 155,577 111,001 44,576 28.6%
Kibondo Mtendeli Burundese 66.080 28.913 37.167 56.2
Kanembwa Burundese 19.560 1S.S31 4.029 20.6
Nduta Burundese , 22.488 13.339 9.149 40.6
Mkugwa Burundese/Rwandese 1.322 1.171 1S1 11.42

SUB-TOTAL (KIBONDO) 109,450 58,954 50,496 46.14%
KIGOMA REGION: GRAND T O T A L 311,092 204,773 106,319 34.17%

2. Chart summary by District

Before Reg. After Reg.
Kigoma Rural Congolese 46.065 34.818
Kasulu Congolese/Burundese 155.577 111.001
Kibondo 8urundese 109.450 58.954

Kigoma Rural Kasulu Kibondo
light shade: indicates population after registration dark shade: indicates population before registration

3. Kigoma region caseload by nationality
Congolese 74.029

Burundese 130.744

Total 204.773

Source: UNHCR, Dar es Salaam

Acceptability and use of cereal-based foods in refugee camps

Appendix 4(d): Market prices at the time of the study

white maize grain bowl 120-140 106
kg 86-93
yellow maize bowl 120
grain kg 86
sorghum bowl 150
millet kg 300

nee kg 450-500 450 450

white maize meal kg 150
yellow maize meal bowl 110-120 110
cassava flour bowl 100 100 100

CSB bowl/kg 90-100 120 n/a

sugar kg 500

small beans bowl 250-300 250

medium beans 350 350 350

old beans** 10-50

lentils 10-50 n/a

green peas bowl 250-300

mahole small/large 100/200 50/100
sweet potatoes heap 100 50/100

yamsxl " 100
irish potatoes small/large 100/200
cassava (fresh) «( 100/200 100

cassava (dried) heap 200 100

bananas bunch 1,700

green vegetable bunch 10
vegetable oil
palm oil litre 700 600

* Amounts are measured in the markets with tin bowls (heaped), which contain around 1.5kg - 1.8kg.
• • Old consignment of beans given earlier this year in the general distribution.

Appendix 4(e): World Food Programme report of
market prices to July 1997


70 40 30 70 40 150 50 150 NA 150
FEB 70 40 30 70 40 150 50 150 NA 200
70 40 30 70 40 150 50 150 NA 200
70 40 30 70 40 150 50 150 NA 200
AV. 70 40 30 70 40 150 50 150 NA 187.5

70 60 30 60 250 180 40 150 NA 200
MARCH 70 60 30 55 250 180 40 150 NA 200
70 60 30 65 250 180 40 150 NA 200
70 60 30 70 250 180 40 150 NA 200
AV. 70 60 30 62.5 250 180 40 150 NA 200

70 50 30 NA 150 180 40 150 NA 200
APRIL 80 40 30 NA 150 180 40 150 NA 200
NA 50 20 NA 150 180 40 150 NA 200
NA 50 20 NA 150 180 40 150 NA 200
AV. 37.5 47.5 25 • NA 150 180 40 150 NA 200

50 60 40 NA 150 180 40 200 NA 200
MAY 50 60 40 NA 150 180 40 200 NA 200
60 50 30 70 100 150 30 250 NA 300
60 50 30 70 100 150 35 300 NA 300
AV. 55 55 35 35 125 165 36 237.5 NA 250

60 30 25 80 30 160 30 400 60 200
JUNE 60 30 25 80 30 160 30 400 60 200
60 30 25 80 30 160 25 400 60 200
60 30 30 80 30 160 25 400 NA 200
AV. 60 30 26.25 80 30 160 27.5 400 45 200

100 20 40 80 20 200 100 400 80 200
JULY 100 20 40 80 20 200 100 400 NA 250
100 35 40 90 120 200 120 400 NA 250
150 40 40 90 150 200 150 400 NA 300
AV. 112.5 28.75 40 85 77.5 200 117.5 400 20 250

NA 70 60 80 150 200 NA 400 NA 250
AUG. NA 70 60 80 200 240 NA 400 NA 300
NA 70 60 90 200 240 NA 400 NA 300
100 90 NA 100 200 250 50 500 NA 300
AV. 25 75 45 87.5 187.5 232.5 12.5 425 NA 287.5




Acceptabilily and use of cereal-based foods in refugee camps

Appendix 4(f): Summary of food ration scales
and kilo calories, January-August, 1997

1 0 350 120 0 20 0 1839
2 400 0 120 30 20 0 2093
3 0 350 0 30 20 5 1551
4 0 175 120 15 0 5 1089
5 200 0 60 0 0 5 901
6 200 175 120 0 20 5 1909
7 200 0 0 30 20 5 991
8 200 0 120 30 20 0 1393
9 0 350 120 0 20 5 1839
10 400 0 120 30 20 0 2093
11 0 350 120 30 20 5 1953
12 400 0 120 30 20 0 2093
13 0 350 120 30 20 0 1953
14 0 350 120 30 20 0 1953
15 0 350 120 30 20 5 1953
16 0 350 120 30 20 5 1953
17 0 350 120 30 20 5 1953
18 400 0 120 30 20 5 2093
19 400 0 120 30 20 5 2093
20 400 0 120 30 20 0 2093
21 400 0 120 30 20 5 2093
22 400 0 120 30 20 5 2093
23 400 0 120 30 20 5 2093
24 400 0 120 30 20 5 2093
25 400 0 120 30 20 5 2093
26 400 0 120 30 20 5 2093
27 400 0 120 30 20 5 2093
28 400 0 120 30 20 5 2093
29 400 0 120 30 20 5 2093
30 400 0 120 30 20 5 2093
31 400 0 120 30 20 5 2093
32 0 350 120 30 20 5 1953
33 0 350 120 30 20 5 1953
34 400 0 120 30 20 5 2093
35 400 0 120 30 20 5 2093



Milling machine

Entry bin for grain / where Maize grain sacks are emptied
Date: 12/9/97

RubbtoM number I
Flour out* Contains maize grao yettobe mfflxti
Grain in
Rdbbhatt nunfcer two/ flout storage

yivii's \ "
\ Sack conveyer / for newly milled flour

KO 6MT's ;nerator
Grain intake box MilUng machine
Flour conveyer / newly baged flour

Rubbhall number I
Grain in rtorage Rubbhall Flour out


Plan by Frank CawkweU, WFP logistics, Kampala
Acceptability and use of cereal-based foods in refugee camps

Appendix 4(h): List of key informants and interviewees
Zaria Saidi, Community Services co-ordinator Lennart Kotsalainen, Acting Representative,
BO, Dar es Salaam
CARE: Richard Ruhogo, Project Officer
Jose Canhandula, Programme Officer,
CHRISTIAN OUTREACH: Susanna Wigram, Dar es Salaam
Community Services co-ordinator/Acting Luc Stevens, Acting HOSO, Kigoma
Country Director, Kasulu Zahra Mirghani, Nutritionist, Kigoma region
Lucas Machibya, Field Assistant (nutrition)
IFRC: Jerry Talbot, Head of'Delegation,
Kigoma region
Dar es Salaam
Bushra Halepote, Field Co-ordinator, Kasulu
Oxfam GB
Modhakkiru Katakweba, EPC Arusha
Irene Lacy, Country Director, Dar es Salaam
Burton Twisa, Programme Manager, Kasulu
Eddie Rowe, Acting HOSO, Kigoma
Theobadina Kasakanta, Community worker,
Bradley Currant, HOSO, Kigoma
Nduta camp
Rashid Hussein, Field Officer, Kasulu
Tanzania Red Cross Society Sixtos Michael, Food Monitor, Muyovosi camp
Dr Mtera, Health Co-ordinator, Kasulu Frank Cawkwell, milling engineer, Isaka
Filimon Malaja, nutrition supervisor, Terry Pinto, Acting HOSO, Isaka
Muyovosi camp
Makanie Issa, clinical officer, Muyovosi camp

Zone Place in family Place of origin (Burundi) DOA1 Direct/via Zaire

1. A Wife ofhlih 3 Mukembe/Rutana Sep'96 Direct
2. A Female hhh (widow) n/a n/a n/a
3. B Male hhh + wife Bujumbura Nov'96 Zaire 4 years
4. B Wife of hhh + son Bujumbura Nov'96 Direct
5. C Wife of hhh Cibitoke Dec'96 Zaire 2 years
6. C Wife of hhh Cibitoke Nov'96 Zaire 2 years
7. C Wife of hhh n/a n/a Zaire 3 years
8. D Wife of hhh Gitega Oct'96 Direct
9. D Wife of hhh
+ sister in law Makamba n/a Direct
10. E Wife hhh
+ widowed friend n/a Nov'96 Zaire 3 years
11. E Wife of hhh Ngozi Nov'96 Direct
12. F Wife ol hhh + friend Gitega Dec'96 Direct
13. F Wife of hhh + 2 friends Bujumbura Nov'96 Zaire 4 months
14. G Wifeofhhh Ngozi n/a Direct
15. G Wife of hhh Ngozi Nov'96 Direct
16. G Wifeofhhh + sister Nyanza lac Nov'96 Direct
17. H Wife of hhh Cibitoke Dec'96 Zaire 3 months
18. H Wifeofhhh + husband Rumonge/Bujumbura Dec'96 Zaire 4 years
19. 1 Wifeofhhh
+ adult daughter Makamba/Nyanza lac Dec'96 Direct
20. 1 Wifeofhhh + husband Cibitoke Jan'97 Zaire 2 years
21- J Wifeofhhh Makamba/Nyanza lac n/a Direct
22. J Wife of hhh Rutana Nov'96 Direct ('90-'93 refugee in TZ)
23. K Wife of hhh Ruyigi Nov'96 Direct
24. K Wifeofhhh + 2 friends Rutana n/a Direct

Date of arrival in Muyovosi camp - Head of household


Nepal case-study gundruk: a dried green vegetable, either radish
leaves, cabbage, mustard, or rape leaves,
achar: pickles made from e.g. mango and ginger which has been processed by hand and left for
which can be stored for a long time. Refers also 10-15 days. It can then be stored for a long
to the spicy mix of tomatoes, chilli, garlic, and time and is used toflavourcurry. It can also be
salt which is made with the low and silotho and used together with other ingredients to make
consumed with rice and daal. Acliar and pickles.
chutney tend to be interchangeable terms, even haldi: turmeric, a yellow powder used for colour-
though chutneys are prepared fresh in ing the tarhari (curried vegetables) It is essential
quantities sufficient only for a meal or two. to the acceptability of curry and is distributed
bagara: parboiled rice (Hindi). In Nepali ushina in the general ration of the refugees.
cliamal. halwa: a food used mpuja. May also be eaten as a
bhat: the Nepali word for all cooked rice, whether special snack. It is made from suji, which is
parboiled or polished. fried mghee. Sugar is added. It is then stirred;
cliamal: the Nepali word for all uncooked rice. water is added and it is brought to the boil,
There are different types of rice, including, for stirring continuously until the water has been
example, chulthe, basmati, and aomusoli (all high absorbed. Usually eaten with a spoon.
quality), and mota (lower quality). household: a group of people who 'share a fire',
cliaulani: the rice water remaining after washing that is, who share food on a regular basis.
the uncooked rice, which is then used for Households usually contain people related by
adding to the tarhari when rice is to be eaten blood or ritual (such as marriage) called families.
without daal. \tdaal is to be eaten, then often People who are not members of the family may
the cliaulani will be thrown out, and the tarhari also live in a household (Scrimshaw, 1987).
eaten after being fried in oil.
janglo: flat woven straw tray on which rice is
chiura:flat,beaten and dried rice, used as a snack placed and then cleaned.
eaten dry or reconstituted in tea or milk or
karai: the frying pan in which vegetables are
fried with ghee. Can be stored for over one
cooked. 'A symbol of domestic auspiciousness
month. Slightly more expensive to buy than
and culinary normalcy'(Khare: 295).
polished rice.
khaja: Nepali word for breakfast, but it can be
daal: collective term for dried, split legumes. The
most common use is as gravy for the main taken at early morning, noon, or mid-
meal. There are different types, including afternoon. A snack. Eating snacks is a regular
chana (yellow), which has been given feature of daily life in Nepal, especially in the
consistently to the refugees for at least two late afternoon and in summer months when
years and is cheaper than masoori daal; halo days are long.
(black), commonly grown in the hills of khole: also called gillo bliat, which means soft
Bhutan but not in Nepal, and therefore rice. Rice is cooked in a large amount of water
cheaper than cliana in Bhutan, but more for 3(M:0 minutes. Salt is added and
expensive than chana in Nepal; masoori (red), sometimes ghee. It is used as a snack, not a
which has been given to the refugees in the meal, and is given to the sick and to children. It
past. Kalo and masoori daal are considered as is more palatable if made from polished rather
'hot' foods and therefore not suitable for dian parboiled rice.
eating in summer or to be given to a sick
person. Other types of daal, and also rice, are low: small stone used with the silotho (see silotho).
not classified either as hot or cold. masala: a ground mixture of cumin, coriander
ghee: clarified butter, used for shallow and deep seed, black cumin, black pepper, ginger, garlic.
frying. mohi: curd thinned with water.

Acceptability and use of cereal-based foods in refugee camps

pauspilho: 'nutritious flour' which was distributed Ethiopia case-study
in Bhutan in schools and health clinics.
Implies flour has been mixed with DSM, oil or aawivwad: donation, gift.
sugar, and is fortified to some extent. alawaat: the kitchen hut.
pitho: a generic term for flour. Used by inter- aqal: nomad's collapsible tent.
viewees to refer to the blended food distributed.
bur: a type of bread, in which the dough is not
prasad: 'Foods, any and all, offered to and fermented. It is poured on the daawa, and
accepted by one's deity of prescription and/or cooked on both sides. Sugar and egg can be
choice. Metaphorically, a material (edible) added to the dough mixture if available.
expression of human and/or divine boon,
blessing, gift, pleasure, favour, disinterested boorash: a generic term for a porridge dish, unlike
kindness, or presence' (Khare:297). the specific shurbad. Commonly used by camp
residents in the study to refer to the porridge
puja: 'Ritual acts — mental or actual — of made out of the ration Famix.
worship, usually in relation to a deity as an
object of worship; worship, reverence, homage, buushe: chaff remaining from the winnowed
or surrender'(Khare: 303). A religious occasion grain, discarded or given to the chickens.
usually organised by the head of the family. cassiset: a porridge made from white flour, the
May be held for a number of reasons and consistency being thicker than that of shurbad.
commonly on the anniversary of the death of a Specifically for newly delivered women.
relative. Rice and other foods are offered to dhananis: the 'starter' added to canjeero dough to
the deity by the pandit (priest). After the begin fermentation.
ceremony is over, the food offerings are taken digir: beans.
by thepandit. The people offering thepuja and
those invited to the ceremony eat selroti and daawa: the circular iron plate (griddle) used for
the making o£ canjeero.
tarkari as prasad.
canjeero (or kimis or laxoox): enjera (the English
puwa: made from rice flour which is fried and a
translation): 'a thin pancake-like sour bread
litde water added and sugar (unlike lialwa,
that is prepared from fermented flours of tef,
which is made from suji with more water added). barley, wheat, corn, sorghum or a combina-
saag: a winter green leafy vegetable (mustard tion of some of these cereals ... the dough used
leaf) widely available and relatively cheap. for making enjera should not be fermented for
Easily grown in vegetable gardens, and the more than three days. It is best to give babies
main crop of the refugees' vegetable gardens. enjera which has been fermented only
It is cooked in oil and water or fried only. overnight, since its nutrient value decreases
Seasoned with fenugreek, garlic and chilli. when it is fermented for a longer period of
satho: maize or wheat which has been fried before time' (EN1 1980). According to informants in
it has been ground (unlikepitho, which has not this study, enjera can be made from wheat flour
been fried). alone, or wheat And faff a, wheat and maize,
selroti: made from rice flour, deep fried in ghee. wheat and sorghum, but not from maize and
sorghum. The main ingredient for Somalis is
Mostly for puja and special occasions.
wheat. The mixture is prepared the previous
silotho: large stone found in all houses for the afternoon or evening and a starter added,
grinding of tomato, chilli, salt, and garlic to which is some of the dough saved from the
make achar. previous fermentation and used for fermen-
sinky: likegundruk, but made from radish. tation of the new dough. Next day, it is poured
suji: semolina, i.e. a wheat flour. More expensive on to the spitting hot daawa in a very thin layer,
dian ordinary wheatflour.Used only for lialwa. covered with a heavy lid for about two minutes,
then removed. It is eaten the same day.
tarkari: cooked vegetables.
Unilito: the blended food product made locally in faxfax: a stew-type meal made from meat mixed
Biratnagar and distributed in the general with cabbage, onion and green tomatoes.
ration and for supplementary feeding, faffa/Famix: the name given to the boorash which is
comprising wheat, maize and soya flour, and given in the ration. It is the trade name of the
fortified with micronutrients. blended food which has been produced in
Addis Ababa, Ethiopia for about 40 years.


garo: can be made from wheat, maize or Tanzania case-study
sorghum. Wheat is cleaned and then boiled;
maize and sorghum have to be pounded to amubese: palm oil.
remove the 'eyes' (germ), then it is boiled. It amaleke (in Kiswahili mahole): a root vegetable
can be eaten with just salt and oil added, or commonly seen in the market.
with milk and butter added, but not with meat
and vegetables. It can be eaten for breakfast, amavuta: oil (any type).
lunch or supper. buswage: a kind of ngali made from pounded,
boiled cassava.
geedo: spice used in cooking.
dona/mar jaune/pate jaune: the name given to
galay (or arabahh): maize.
yellow maize flour (coming pre-packed from
hadhuudh yaryare: sorgh u m. USA). Although dona is a Kiswahili term which
jejeb: green split pea used as filling for samosas. strictly means any maize flour which has not
karis: can be made from wheat or sorghum, not been first pounded before being milled, the
maize. The grain is cleaned and pounded to yellow flour was not known to the refugees
remove the husks, then boiled until soft. Oil, before coming to Burundi (and in some cases,
salt and spices are usually added. Eaten with Zaire before Burundi), so they have no pre-
milk or with meat and vegetables. existing term in Kirundi for this specific
commodity. Hence a variety of terms seemed
kildhi: kettle for boiling water for tea. to be in usage, roughly translated from French
niasaf: straw circular tray for winnowing wheat to 'yellow ugali'.
clean it and separate the enjera from the shura.
ibhibigori: maize grain.
musaqo: maize, sorghum or wheat (cereal).
ibhiliarage: beans (any type).
mooye: the wooden receptacle for hand-pounding
of grain, used with the lip. ibiloke: generic term for plantains/bananas.
quad (English version usually written as 'khatt' or ifu iyela: white flour (any type).
'chat'): the green leaves chewed by men for ifuyu musululu: flour for porridge.
recreation, with a mild amphetamine-type imyumbati: cassava plant/root.
effect. indagara: small fish from Lake Tanganyika.
qarde: credit obtained from a shop or an drink made from sorghum,
individual. which becomes alcoholic after fermentation. A
shurbad: porridge made with wheat, sorghum or typical traditional low-alcohol, opaque beer.
maize flour with sugar, milk and butterfat inkonge/mahonda: sorghum, red or white.
added. Traditionally made for women who
insekulo: wooden base for pounding grain, used
were newly delivered and children under 5
with the muhini.
years. Also a fasting food (Ramadan).
shiide: manual grinding stone. intengwa/incarum (in Kiswahili mbaazj): small white
bean cultivated in Tanzania, especially Kigoma
shuuro: a commonly made dish, usually for region, and in some of the refugee gardens.
supper. Made from the large particles of
ground wheat, maize or sorghum. This is ishiga: the round, fuel-efficient stove.
boiled with water, stirring continuously. kazemukangara: a type of bean which refugees
Before removing from the heat, oil is stirred in used to cultivate in Burundi.
and enjera flour may be added to thicken the Itambaranga: the name given by the refugees to
consistency (which is like that ofugali). Eaten
the old beans received in the ration.
'dry', but preferably with milk or with meat
and vegetables. lengalenga/mbwiga (in Kiswahili mchicha): a green
lip: the wooden instrument for pounding grain in spinach-type vegetable.
the mooye. mashiga: three-brick/stone stove.
xash: wheat flour/Famix with oil and sugar muhini: wooden stick for pounding in the insekulo.
(premix). musulaki: musulac is a low-cost blended comple-
xasiid: new maize which is roasted. mentary food, locally produced in Burundi,
the production assisted by donors. The
(Sofia Abdi Ahmed, of Hartisheihh A camp, helped to
product has been promoted through a social
compile the Ethiopia glossary.) marketing programme

Acceptability and use of cereal-based foods in refugee camps

mutobe/amake: drink made from bananas. ubunyange: cassava flour (white), prepared by first
ndete: a meal made from pounded maize, boiled soaking the peeled cassava for 3-7 days, then
with beans or lentils, with salt and oil added. drying it, then grinding it.
ntole (Kiswahili), nyanya mshumaa (Kirundi): a ubuzyenge: cassava flour (off-white) which has
green cucumber-type vegetable commonly been prepared by peeling, covering for two
eaten widi ugali. days with banana leaves to ferment, then
drying and grinding.
ubugali (in Kiswahili ugali): very stiff porridge
eaten with meat, vegetables,fish,beans. Made ugwagwa: alcoholic drink made from bananas.
from maize or cassava flour, a main staple dish. umusululu (in Kiswahili uji): thin porridge of any
ubulo (in Kiswahili ulezi): millet (used for porridge type.
after grinding and also in the production of
alcohol and can be usedforugali).


Abdulkadir, A. (1994) 'A Survey on the FAO (1995) 'Sorghum and Millets in Human
Repatriation of Somali Refugees to Northern Nutrition', FAO, Rome.
Somalia from Eastern Ethiopia'. Farah A. Y. (1994a) 'Profile of the Camp
ACC/SCN (1995) Report of a workshop on the Populations in the Somali Region 5
Improvement of the Nutrition of Refugees and (E. Ethiopia)', SCF-UK/ARRA.
Displaced People in Africa, Machakos, Kenya, Farah, A. Y. (1994b) 'Determinatives ofa Large
5-7 December 1994. United Nations Admini- Scale Return-home Movement: Integration of
strative Committee on Coordination — Camp Dwellers in Eastern Hararghe', SCF-
Subcommittee on Nutrition with University of UK/ARRA.
Nairobi, Applied Nutrition Programme (ANP). Gautam, M. (1993) 'Nutritional Assessment of
Ambroso, G. (1994) 'The Somali Clan System: an WFP Food Ration for Bhutanese Refugees in
Introduction to Somali Society and History', Nepal', WFP, Nepal.
UNHCR Ethiopia. Gittelsohn, J. (1989) 'Intrahousehold Food
ARRA/UNHCR (1996) Joint nutrition survey, Distribution in Rural Nepal', PhD
Eastern refugee camps, May 1996. dissertation, University of Connecticut.
Baker J. K (1995) 'Nutritional Status and Gladwin, J. (1993) 'Results of a Field Trial of
Nutritional Disorders in Bhutanese Refugees Instant Blended Food Produced by the Milk
Residing in Nepal', MSc dissertation, University Marketing Board, UK, Used on Supple-
of Cambridge. mentary Feeding Programmes for Moderately
Bal Kumar, K. C, and Prakash dev Pant (1996) Malnourished Children in Rohingya Refugee
'Socio-economic and Demographic Survey of Camps, Cox's Bazar, Bangladesh September
Refugees from Bhutan in Eastern Nepal, 1995', 1992 to February 1993', Liverpool School of
Tribhuvan University, Kathmandu: The Tropical Medicine.
Central Department of Population Studies. Golden, M.H.N. (1991) 'The nature of
Beaton, G. H. (1995) 'Fortification of Foods for nutritional deficiency in relation to growth
Refugee Feeding', Canadian International failure and poverty', Ada Paediatr Scand Suppl
Development Agency. 1991; 374:95-110.
Bhatia, R. (1996) Review of Food and Nutrition Halepote, B. (1997) personal communication, 1997.
Programmes, Great Lakes, 24 March-18 April Hansch, S. (1992) 'Diet and ration use in Central
1996. UNHCR Geneva, PTSS. American refugee camps', Journal of Refugee.
Bhatia, R. (1997) personal communication, Studies 1992; 5:3/4:300-12.
December 1997. Henry, C.J.K. and J. Seaman (1992) 'The micro-
Briend, A. (1995) "Blended food" in 'Report ofa nutrient fortification of refugee rations to
Workshop on the Improvement of the prevent nutritional deficiencies in refugee diets',
Nutrition of Refugees and Displaced People in Journal of Refugee Studies 1992; 5:3/4:359-67.
Africa, Machakos, Kenya, 5-7 December Hogg, R. (1992) 'Changing Mandates in the
1994', ACC/SCN. Ogaden: the Impact of Somali Refugees/
Buggeland, A. (1993) 'Social Survey of Leader- Returnees on the UNHCR', UNRISD.
ship Structure in Bhutanese Camps in Nepal', Hurt, M. (1994) Perspectives on Conflict and Dissent
UNHCR, Nepal. (Bhutan), Kiscadale Asia Research Series No.4,
Cawkwell, F. (1997) personal communication, Gartmore, Scotland: Paul Strachan-Kiscadale.
September 1997. Joyce-Jenkins, A., B. Rushatsi and E. Ngezi
Clugston, G. A. (1994) 'A Review of Nutrition and (1996), 'Nutritional Survey of Kasulo and
Food Situation amongst the Bhutanese Mwivuza Villages, Kagera Tanzania, Nov/Dec
Refugee Camp Populations' (Draft report), 1996.'
WHO, Geneva. Keen, D. (1992) Refugees: Rationing the Right to
DHA (1997) IRIN (Integrated Regional Life, London and New Jersey: Zed Books.
Information Network) for the Great Lakes Khare, R.S. (1976) The Hindu Hearth and Home,
update report. Durham: Carolina Academic Press.

Acceptability and use of cereal-based foods in refugee camps

Lacy, I. (1997) personal communication, 1997. Shrestha, D.L. and Ganai Tako (1996) An
Majupuria, I. (1980) Joys of Nepalese Cooking, Overview of Water Induced Hazards and
Jullundur: Smt. S. Devi. Mitigation Measures in Bhutanese Refugee
Marrack, C. (1994) 'Review of the Bhutanese Camps in Nepal', UNHCR, Nepal.
Refugee Health Programme, Jhapa District, Sommers, M. (1995) 'Representing refugees: the
Nepal', SCF-UK. role of elites in Burundi refugee society',
Miles, M.B. and M.A. Huberman (1994) Disasters 19:1:19-25.
Qualitative Data, Analysis, Thousand Oaks, SUSTAIN (1997) 'Results Report on the Vitamin
London, New Delhi: Sage Publications. C Pilot Program', Washington DC.
Ms. Halimo (1997) personal communication, Thacker, P. and B.M Sharma (1996) Socio-
June 1997. economic and Environmental Impact of the
New ERA (1993) 'Mitigation of the Impact of Presence of Bhutanese Refugees on the Local
Bhutanese Refugees on Local Communities', Community in Eastern Nepal (A situational
LWS Nepal. analysis)', UNHCR, Nepal.
Nieburg, P., B. Person-Karell and M. J. Toole Thiadens, R. (1996) 'Tanzania: Technical Mission
(1992) 'Malnutrition-mortality relationships in the Agriculture Sector 14-24 September
among refugees', Journal of Refugee Studies 1996', UNHCR, Geneva, PTSS Mission Report.
5:3/4:247-56. Toole, M.J. (1993) 'Protecting refugees' nutrition
Oxfam UK and Ireland (1996) 'Oxfam Jijiga with food aid', Nutritional Issues in Food Aid: ACCI
ETH 242 Programme Evaluation'. SCN Symposium Report, Nutrition Policy Discus-
Robertson, J. (1994) 'Nutrition Survey Report of sion Paper no. 12, Geneva: ACC/SCN, 1993.
an Investigation of Suspected Beri Beri in the Toole, M.J. and R. Bhatia (1991) Case Study:
Bhutanese Refugee Programme, Nepal', SCF-UK. Somali refugees in Hartisheik A Camp,
Savage King, F. and A. Burgess (1992) Nutrition Eastern Ethiopia', Health and nutritional
for Developing Countries, Vol. ELBS 2nd ed., profile, July 1988-June 1990.
Oxford: Oxford University Press. TRCS (Tanzania Red Cross Society) (1997)
SCF Dire Dawa( 1988)'Short Report of Nutrition 'Health Sector— Monthly Report', Muyovosi,
Survey of the 5 Ogaden Awrajas and Jijiga August 1997.
Awraja'. Tsegaye, W. (1996) Report on Supplementary
SCF-UK. (1991) 'An End to Isolation: the Report Feeding Programme, ARRA.
of the Ogaden Needs Assessment Study', Tsegaye, W. (1997) personal communication,
London: SCF-UK. May 1997.
SCF-UK (1996) 'Household Food Economy UNHCR (1997) 'Commodity Distribution: A
Assessment of Somali Refugee Camps Region Practical Guide for Field Staff, UNHCR, Geneva.
5, Eastern Ethiopia'. UNHCR Kathmandu (1997) Briefing note,
SCF-UK Jijiga (1991) 'Activity Report: January 1997: 'Bhutanese Refugees and
Hartisheikh Refugee Camps Ethiopia, Asylum Seekers in Nepal', UNHCR, Nepal.
September-December 1990'. UNHCRJhapa (1994a) 'Acceptability Survey of
SCF-UK Jijiga (1996) 'Integrated Food Security Blended Food/Unilito During a Distribution
Project, Somali National Regional State in Beldangi 1 Refugee Camp', UNHCR,
(Region 5)', project proposal. Nepal.
Scrimshaw, S. (1990) 'Combining quantitative UNHCRJhapa (1994b) 'A Study on the Outflow
and qualitative methods in the study of intra- of Relief Commodities from the Bhutanese
household resource allocation' in B.L. Rogers Refugee Camps', UNHCR, Nepal.
and N.P. Schlossman (eds.): Intra-lwuseliokl UNICEF/TFNC (1997) 'Report on Nutrition
Resource Allocation, Food and Nutrition Bulletin. Survey in Mtendeli and Mtabila Extension
Tokyo: United Nations University Press, 1990. Camps in Kigoma Region, Tanzania
Scrimshaw, S. and E. Hurtado (1987) Rapid September 1997' (draft only).
Assessment Procedures for Nutrition and PrimaryUpadhyayJ.K. (1994) 'Blended Food: Report of
Health Care: Anthropological Approaches to a Mission to Nepal, 3 Jan to 19 Jan 1994',
Improving Programme Effectiveness, Tokyo: UNHCR Geneva.
United Nations University, UNICEF, and Upadhyay, J.K. (1997) personal communication,
UCLA Latin American Center. 1997.
Shoham J., J. Rivers and P. Payne (1989) van Brabant, K. (1994) 'Bad Borders Make Bad
'Hartisheikh', London School of Hygiene and Neighbours: the Political Economy of Relief
Tropical Medicine. and Rehabilitation in the Somali Region 5,


Eastern Ethiopia', Relief and Rehabilitation WFP/UNHCR (1996b) 'Joint WFP/UNHCR
Network, OD1: London. Food Assessment Mission in the Great Lakes
WFP (1995) 'An Analysis of the Food Security Region (local mission).'
Situation in Burundi.' WFP/UNHCR (1997) Memorandum of Under-
WFP (1997) 'Project Ethiopia 5241 (Exp.2)', standing on the Joint Working Arrangements
WFP, Ethiopia. for Refugee, Returnee and Internally
WFP/UNHCR (1996a) 'Ethiopia: Joint Displaced Feeding Operations.
WFP/UNHCR Donor Food Aid Assessment Wikan, U. (1990) 'The Girl Child in Bhutan',
Mission,June 1996.' UNICEF/University ofOslo.

Order Form

Acceptability and Use Name
of Cereal-Based Foods
in Refugee Camps Address

(154024) £12.95 | $18.95 each

Country Postcode/Zip
sub-total: £
Address for delivery
(if different from above)
Total value of order: £

* Postage and packing: £

Total payment: £

I enclose payment by cheque/PO ''Postage and Packing:
(customers from N America should make
cheques payable to Humanities Press; all other UK and EU orders: add 20% of order value up to a
cheques should be made payable to Oxfani) maximum p&p of £ 15.00. Please send your order to
the UK postal address given below. All orders from
Please accept payment by the individuals must be prepaid.
following method (please circle) All other overseas orders excluding North America:
Please add 30% of order value for postage and packing
Access / Mastercard / Barclaycard costs up to a maximum p&p of £40.00. Please send
Visa / American Express your order to the UK postal address given below.
(non-US customers only) Orders from individuals must be prepaid.
Ordersfromthe USA and Canada: Orders from
Card number individuals must be prepaid. All such orders will be
supplied by UPS at the additional charge of $4 for the
Expiry date first book and $.75 for each additional book ordered,
to cover postage and handling. Please supply street
address and send your order to the US postal address
given below.

Customers in North America All other orders should be sent to:
should send their order to:
Oxfam c/o Humanities Press c/o BEBC, PO Box 1496
165 First Avenue, Atlantic Highlands Parkstone
NJ 07715-1289, USA Dorset BH12 3YD, UK
telephone 732 872 1441 telephone 01202 715555
fox 732 872 0717 fax 01202 715556


Master your semester with Scribd & The New York Times

Special offer for students: Only $4.99/month.

Master your semester with Scribd & The New York Times

Cancel anytime.