31. K , .
. 13.5.3

gr
. balm”?
, . hum

 

 

‘ .1 ‘

VA. . ,-l4..... (IE \
.
.

wfifihwfluw. .

2. .3. .920 .‘
. 4:. .IV ,2...

 

 

I

.f? 7'97

 

LIBRARY
IVIIQI "gal I State
University

 

 

 

This is to certify that the
dissertation entitled

DIAGNOSTIC STUDY OF FARMERS SEED QUALITY AND
VARIETY PREFERENCE IN SOUTHERN MALI

presented by

Marthe Diallo

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Crop and Soil Science

 

 

 

/ wMaKJr Professor's'Signature
CDT/n e / 01 200?

 

Date

MSU is an Affirmative Action/Equal Opportunity Employer

4 -—-—.----

A—.-.—-.--.-—4-.—.-.----—.A

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K lPrQ/Acwpres/CIRC/DateDue Indd

 

DIAGNOSTIC STUDY OF FARMERS SEED QUALITY AND VARIETY
PREFERENCE IN SOUTHERN MALI

By

Marthe Diallo

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Crop and Soil Sciences

2009

ABSTRACT
DIAGNOSTIC STUDY OF FARMERS SEED QUALITY AND VARIETY
PREFERENCE IN SOUTHERN MALI
By
Marthe Diallo

Mali relies on agriculture as the backbone of its economic growth because
agriculture employs more than 80% of the p0pulation and constitutes the main source of
food. In the Koulikoro region (the study area), sorghum and millet are subsistence crops
for almost all of the population. Groundnut constitutes the main source of income for
women and is critical to family nutrition. These crops are grown by almost all the farmers
in this region because they are adapted to semi-arid tropical ecologies and infertile soils,
as well as being preferred in the diet. Despite their importance, an understanding of seed
systems for these ‘orphan’ crops has been neglected.

To understand agricultural development in Mali, it is important to evaluate the
entire seed system, including the seed quality of farmer-saved and foundation seed, seed
saving methods on-farm, seed sourcing avenues, and how new varieties are accessed by
small-holders. Informal and formal surveys were conducted, in conjunction with seed
collection from a range of sources. Seed quality performance was tested for seed from a
range of sources. To quantify seed quality, laboratory standard analytical methods and
field performance were analyzed. Variety purity was also assessed for different seed
sources that represented foundation seeds, and seeds maintained by farmers for different
number of years. These tests were linked to farmer assessment of seed quality and variety

preference conducted through surveys and farmer ranking exercises.

The seed quality assessment showed that farmer produced seed is generally high
quality, as high as foundation seed. The seed meets Mali national seed service standards
in almost all cases. However, there were some concerns identified concerning physical
purity and health status of groundnut seeds.

The field trial analyses showed that there are no significant differences between
the variety purity of seeds farmers had saved and produced and foundation seed. This was
shown for the flowering time (beginning and end), number of off- types, and weight of
panicles. Overall, farmers recycling of varieties (saving seed for several years) did not
markedly alter variety traits in sorghum.

Results from both the on-farm field trial evaluation and the survey indicated that
yield and adaptation to the local environment were very important in farmers’ variety
evaluation criteria. There were some additional traits of interest to farmers from Dioila
and Mande, including cooking traits (easy processing, good taste of dishes made) and
drought tolerance. However the farmers in Dioila were more interested in cooking quality
than in drought tolerance.

Another interesting finding is that there is a cultural tradition that prohibits the
purchase of seed, particularly for sorghum and groundnut, yet farmers’ did express a
willingness to pay for sorghum seeds of preferred, improved varieties. Farmers in both
zones were ready to pay 200 FCFA/kg for seed for their preferred varieties. This is a
surprisingly high price given the low income level, and the limited cultural concepts of

paying for subsistence crop seed.

 

DEDICATION
To the loving memory of my mother Man’am Sanogo who did not live to see the
accomplishment that built the foundation.
To my husband Nyadia and my two kids Esther and John—Ariel Goita for their

love, support and patience when it was most needed.

iv

ACKNOWLEDGEMENTS

Undertaking graduate studies in the US by a Malian woman would not have been
possible without the assistance, guidance and support by a number of people and
institutions. Many people provided inputs in various aspects of my study. I give thanks to
all who have been involved, several of whom I must mention by name.

First, my foremost acknowledgements goes to my promoters, Dr John .M. Staatz
and Dr Nangho Dembe'lé, for their unfailing support and encouragement in various
aspects of this study at Michigan State University (MSU).

I am extremely grateful to the farmers in Dioila and Mande who participated in
the research of my dissertation. Without their collaboration this research could not be
conducted. I would like to thank ICRISAT- Mali office especially the “programme
sorgho” team. Dr Rattundes Eva and her husband Fred and all their team for war
welcome and all their helps in materials and staff for conducting the field work. I am
thankful to Sidy Dembele (ICRISAT), Arouna Sangare (ICRISAT), Karim Cisse (IER),
Bocar Diallo (IER), Abdoulaye Sangare (ACOD), Mamadou Coulibaly (AOPP),
Mamoutou Diarra (ULPC), Ibrahim Camara (OHVN), Moussa Kanoute (ICRISAT),
Adama Birama Diakite, Deborah Sanou, Sabine Togola for all their helps in the data
collection.

I owe thanks to all my committee members: Dr Sieglinde .S. Snapp my major
professor Dr Russell Freed both of the Department of Crop and Soil Science department
Dr John M Staatz of Department Agricultural, Food, and Resource Economics, and Dr
Eva Weltzien Rattunde of ICRISAT-Mali for the scientific rigor they showed during the

different steps of elaboration of my dissertation . Their sincere interest in this work and

constant availability to discuss research issues will always be remembered. To all of you
my deepest gratitude for making my dream becomes a reality.

A I would also like to thank USAID- Mali office for financing part of my studies
and field trips back in Mali. I really appreciate your true AID for the Malian people by
contributing to human resource capacity building through training.

I am also gratefiil to the Borlaug Leadership in Agriculture Program (LEAP) for
funding the field work in Mali of my dissertation. In the same way I am thankful to the
McKnight foundation for one year assistantship funding that I used to write my
dissertation.

I am grateful to all the faculties and staff of crop and soil science department for
all their help from the course-work to the dissertation writing. I am want thank namely
Dr Lawrence Copland, Dr James Kelly, Dr Sasha, Dr Freed, Calvin Bricker, Rita House,
Darlene Johnson for their constant availability and the wonderful ideas they provided for
the success of this dissertation.

I am grateful to Dr Valerie Kelly, Associate Professor in the Department
Agricultural, Food, and Resotrrce Economics of MSU, for her valuable inputs in editing

and formatting my dissertation.

I have also benefited greatly from the guidance and support from families, fiiends,
and colleagues from Mali. I am thankful to my family in-law for their war welcome and
all their support during the filed work. I am especially grateful to Chitan Keita called

Batoma for her help with baby. I would like for my bothers and sisters to use this work as

an example to follow and even to do more than I have done.

vi

Finally I am thankful to my fiiends: the Longabaugh family, Tomokazu Nagai,
Tracy Breedy, the Wolf and Rose family, and Katjiuongua Hikuepi, Mme Dieng F atou
Faye, F atoumata Dembele, Sara Sogoba the Diarra family, in Houston Texas, all the
African students at MSU for all their support during the writing of my dissertation. I
thank my church family, The International Fellowship and Friends at Trinity Church, for
the constant moral and emotional support.

This list is long but perhaps not exhaustive. As an African proverb says:
“Knowledge is like a baobad tree — one person '5' arms are not enough to encompass it ”;
if by mistake I omitted someone’s name here, I offer all my apologies. It is not that
she/he is not important or his /her help have been underestimated, but these are the names

I could think of now.

vii

TABLE OF CONTENTS

 

 

LIST OF TABLES ................................................................................ xi
LIST OF FIGURES .............................................................................. xvii
LIST OF ABREVIATIONS ................................................................... xviii
INTRODUCTION 1
OVERVIEW OF SORGHUM, PEARL MILLET AND GROUNDNUT ........................... 6
Sorghum and Pearl millet ........................................................................................ 6
Groundnut ................................................................................................................ 7
STUDY AREAS .................................................................................................................. 8
ICRISAT Research Station in Samako (Mali ......................................................... 9
Study zones ............................................................................................................ lO
Brief Overview of Agriculture in the Study Zones ............................................... ll
Socio-Economic and Cultural Situation of the Research Zones ........................... l2
CHART ER 1
QUALITY OF FARMER PRODUCED SEEDS OF SORGHUM, GROUNDNUT,
AND PEARLMILLET IN DIOILA AND MANDE ZONES ..... 15
INTRODUCTION ............................................................................................................. 15
Farmers’ seed sourcing in Mali .............................................................................. l6
Farmers’ local seed management for sorghum, groundnut and millet in southern
Mali .......................................................................................... 19
Biological differences between cereal crops and oilseeds... Error! Bookmark not
defined.
Seed quality parameter assessed in this chapter ....Error! Bookmark not defined.
Research Objectives ............................................................................................... 24
Research Hypotheses ............................................................................................ 24
MATERIAL AND METHODS ......................................................................................... 26
Choice of the research areas ................................................................................. 26
Field Activities ...................................................................................................... 27
Data Collection Methods Used .............................................................................. 28
Data Documentation and Analysis ........................................................................ 35
RESULTS .......................................................................................................................... 37
Surveys ................................................................................................................... 37
Lab analyses .......................................................................................................... 44
Farmers seed storage fi'om survey ......................................................................... 54
Farmers’ assessment of seed quality ..................................................................... 57
DISCUSSION .................................................................................................................... 59
Cropping system characterization ......................................................................... 59

viii

Farmers’ seed sourcing ......................................................................................... 60

Seed storage ........................................................................................................... 61
Seed Quality ........................................................................................................... 62
CONCLUSION ................................................................................................................. 69
CHAPTER 2
IMPACT OF FARMERS’ SEED RECYCLING ON UNIFORMITY OF
SORGHUM VARIETIES IN SOUTHERN MALI ........................................ 70
INTRODUCTION ............................................................................................................. 70
Biodiversity and farmer’s selection criteria .......................................................... 71
Seed industry requirements for uniformity ............................................................ 72
Sorghum ................................................................................................................. 73
Objective ................................................................................................................ 75
MATERIALS AND METHODS ....................................................................................... 76
Research Activities ................................................................................................ 76
Surveys ................................................................................................................... 76
Assessment Varietal Purity or Homogeneity: Field Trial ...................................... 77
Data documentation and analysis .......................................................................... 81
RESULTS .......................................................................................................................... 83
Results from surveys ............................................................................................. 83
Field trial ................................................................................................................ 84
DISCUSSION .................................................................................................................... 93
Farmers’ diversity management ............................................................................ 93
Variety Mixture or Misidentification by Farmers .................................................. 93
Field Uniformity Experiment ................................................................................ 94
CONCLUSION ................................................................................................................. 98
CHAPTER 3
SORGHUM VARIETY TRAIT EVALUATION WITH FARMERS IN
SOUTHERN MALI ............................................................................... 99
INTRODUCTION ............................................................................................................. 99
Research Objectives ............................................................................................. 102
Hypotheses ........................................................................................................... 102
MATERIALS AND METHODS ..................................................................................... 103
Participatory variety selection trials .................................................................... 103
Farmer evaluation ............................................................................................... 104
Samples ................................................................................................................ 106
General farmers’ survey ...................................................................................... 107
Data documentation and analyses ....................................................................... 107
RESULTS ........................................................................................................................ 109
On-farm field trial evaluation by farmers ............................................................ 109
Gender considerations in varietal trait evaluation ............................................... 112
Farmer survey ...................................................................................................... l 13
Results from farmers’ survey per zone ................................................................ l 14

ix

DISCUSSION .................................................................................................................. l 16

CONCLUSIONS ............................................................................................................ 120
CHAPTER 4
IMPROVING FARMERS SEED PROVISION FOR SORGHUM .................... 121
INTRODUCTION ........................................................................................................... 121
Sorghum seed marketing in Mali ........................................................................ 123
Seed fairs ............................................................................................................. 124
Willingness to pay ................................................................................................ 125
Objectives ........................................................................................................... l 25
Hypotheses ........................................................................................................... 126
MATERIAL AND METHODS ....................................................................................... 127
Seed fair survey .................................................................................................... 127
Willingness to pay survey ................................................................................... 128
General farmers’ survey ...................................................................................... 129
Data documentation and analyses ....................................................................... 129
RESULTS ........................................................................................................................ 131
Results from the general survey ........................................................................... 131
Result from farmers’ willingness to pay survey ................................................. 133
Evaluation of the seed fairs ................................................................................. 136
DISCUSSION .................................................................................................................. 144
Willingness to pay assessment ............................................................................ 145
Seed fair evaluation ............................................................................................. 148
CONCLUSION ............................................................................................................... 150
APPENDIX ........................................................................................ 152
Appendix 1 ................................................................................. 152
Appendix 2 ................................................................................. 153
Appendix 3: questionnaire for seed fairs survey ...................................... 154
Appendix 4: Seed samples collection questionnair ................................... 156
Appendix 5: Questionnaire for variety evaluation 2007 in Dioila and Mande
zones ................................................................................................ 157
Appendix 6: Heads of the household interviews ..................................... 158
Appendix 7: Questionnaire individuel pour l’arachide .............................. 159
Appendix 8: Questionnaire individuel pour le sorgho et petit mil .................. 164
Appendix 9: grouping of the farmers variety traits preferences .................... 170
Appendix 10: List of codes .............................................................. 171
Appendix 11: Survey consent statement ............................................... 176
REFERENCES 177

 

 

LIST OF TABLES

Table 0—1: Socio-Agro-Ecological conditions of Dioila vs. Mande ......................... 11

Table 1-1: Seed quality parameter characteristics based on seed certification standards
used by the International Seed Testing Association (ISTA) and the “Comite Inter-état de
Lutte contre la Sécheresse au Sahel” (CILSS) ................................................... 23

Table 1-2: Field activities conducted in the two zones: Dioila and Mande where the
research was conducted from May to December 2007 ......................................... 28

Table 1-3: The seed quality measurement parameters used for the laboratory analyses
conducted over the period October-December 2007 on seed samples collected from
farmers in Dioila and Mande, Mali in May 2007 ............................................... 31

Table 1-4: General information (gender, age, information on the field) of the farmers who
participated in the research in 2007 in the two zones: Dioila and Mande .................. 39

Table 1-5: Information on the participating household cropping systems in Dioila and
Mande ................................................................................................. 40

Table 1-6: Seed samples collected from farmers in Dioila and Mande on May 2007 for
lab-analyses of farmers’ seed quality parameters .............................................. 40

Table 1-7: The number of seed samples collected in eight villages representing two zones
in Mali; Dioila and Mande ......................................................................... 41

Table 1-8: Number of seed samples collected per gender and per crop (groundnut, pearl
millet, and sorghum) in Dioila and Mande for farmers’ seed quality assessment study...42

Table 1-9: Number of seed samples collected per type of varieties (modern versus local)
for all three crops in Dioila and Mande .......................................................... 42

Table 1-10: Number of seed samples collected per seed sources for all three crops
(groundnut, pearl millet and sorghum) in both zones: Dioila and Mande ................... 43

Table 1-11: Percentage of seed samples sourced in market versus farmers own produced
seed per zone (Dioila and Mande) and per crop type (groundnut, pearl millet and
sorghum) ............................................................................................. 43

Table 1-12: Number of seed samples collected, by seed storage method for sorghum,
millet and groundnut per zone ..................................................................... 44 _

xi

Table 1-13: Seed quality characteristics are presented for modern and local sorghum
varieties, along with standard Error of the mean (SE) and a GLM contrast to test if
modern versus local varieties are significantly different ....................................... 47

Table 1-14: Seed quality characteristics for Dioila and Mande sorghum varieties, along
with standard error of the mean (SE) and a GLM contrast to test if quality characteristics
of seed from Dioila are significantly different from those from Mande ..................... 48

Table 1-15: Seed quality analyses results for sorghum are presented by village of
production Mean, and Standard Error presented in parentheses, to identify which villages
have better quality seed ............................................................................. 48

Table 1-16: Factors contributing to high values in impurities observed among sorghum
seed samples, including varieties and villages where seed were produced .................. 49

Table 1-17: Seed quality characteristics for Dioila and Mande for pearl millet and
groundnut, along with standard Error of the mean (SE) and a GLM contrast to test if seed
from Dioila and Mande are significantly different ............................................. 49

Table 1-18: Seed quality characteristics per storage methods commonly used by farmers
for each crop, along with Standard Error (SE) and a GLM contrast to test if quality
measures for seeds stored in panicles, sheaves, and grain are significantly different. . ....50

Table 1—19: Seed quality characteristics per storage methods mainly used by farmers for
groundnut and pearl millet, along with standard error of the mean (SE) and a GLM
contrast to test if seeds stored by different method are significantly different .............. 52

Table 1-20: Seed quality characteristics are for sorghum, pearl millet and groundnut seed
samples along with standard error of the mean (SE) and a GLM contrast to test if seed of
the three crops are significantly different from each other .................................... 52

Table 1-21: Seed quality characteristics per gender for groundnut seeds collected from
farmers in Dioila and Mande zones along with standard error of the mean (SE) and a
GLM contrast to test if seed from men and women are significantly different ............ 53

Table 1-22: Seeds storage method used by farmers in Dioila and Mande, for groundnut,
millet, and sorghum observed during farmers seed quality assessment study done from
May to December 2007 in southern Mali ....................................................... 54

Table] -23: Factors of concern for farmers during seed storage of the groundnut, millet,
and sorghum in Dioila and Mande identified during the farmers seed quality assessment
study done from May to December 2007 in southern Mali ................................... 55

Table 1-24: Seed treatment used by farmers in Dioila and Mande during seed storage for

the groundnut, millet, and sorghum documented during farmers seed quality assessment
study done in 2007 in southern Mali ............................................................ 56

xii

Tablel-25: Farmers’ seed quality evaluation as assessed through survey questions
conducted before and after planting on December 2007 in the eight villages representing
Dioila and Mande zones in Mali .................................................................. 57

Table 1-26 shows Number of seed samples collected per village and per crop type for the
seed quality analyses in the laboratory on 2007 in Dioila and Mande zones .............. 152

Table 1-27 shows the sample size of head of household interviewed during the survey
for seed quality assessment on 2007 study per zone .......................................... 153

Table 2-1 : Places and approximate dates of the different activities undertaken to conduct
the field trial assessing the impact of farmers’ seed recycling on sorghum variety trait
uniformity from May to December 2007 in ICRISAT-Mali research station ............... 76

Table 2-2: Description of the four sorghum varieties used for the field trial fi'om the
catalog of sorghum varieties in Mali obtained with ICRISAT research technician. . . . . ...79

Table 2-3: Number of seed samples (recycling categories and seed categories according
to the source) used in the field for each of the four varieties. Field trial conducted on
ICRISAT-Mali research station on 2007 ......................................................... 80

Table 2-4: Description of the GLIMMEX model the statistical model used to analyze the
field trial data ........................................................................................ 82

Table 2-5: Number of seed samples collected from farmers in Dioila and Mande zone for
the field trial conducted on the ICRISAT research station from July to December 2007 in
Mali ................................................................................................... 83

Table 2-6: Number of sorghum varieties grown by farmers from farmers’ survey
conducted on December 2007 ..................................................................... 84

Table 2-7: Results of GLIMMIX Procedure for flowering characteristics of the field trial

data, including the fixed effects of varieties, recycling categories and recycling categories
within variety and LSM comparisons among varieties and recycling categories, presented
as Julian days ........................................................................................ 85

Table 2—8: Results of LSM comparisons of recycling categories within varieties for
flowering response variables (means, with standard errors in parentheses) ................ 86

Table 2-9: Results of GLIMMIX Procedure for plant phenotypic traits (plant height (cm)
and number of off-type plants for varieties, recycling categories alone and recycling
categories within variety and LS mean comparisons among varieties and recycling
categories ............................................................................................. 87

xiii

Table 2-10: Results of LSM comparisons of recycling categories within varieties for plant
phenotypic response variables (plant height and number of off-type plants) (Means, with
standard errors in parentheses) ..................................................................... 88

Table 2-11: Results of GLIMMIX Procedure for plant yield components (panicle yield
and grain yield (fixed effects) of varieties, recycling categories alone and recycling
categories within variety and LS mean comparisons among varieties and recycling
categories ............................................................................................. 89

Table 2-12: Results of LSM comparisons of recycling categories within varieties for plant
yield response variables (panicle yield and grain yield) (Means, with standard errors in
parentheses) .......................................................................................... 91

Table 2-13: Farmers genetic diversity management for sorghum assessed by recording
the duration of farmer growing the same variety and the percentage of farmers who
reported of having change their preferred dominant variety at least one time during
farmer general surveys conducted in Dioila and Mande zones ............................... 91

Table 3-1: Name of the 32 varieties that farmers evaluated to assess farmers variety trait
preference in December 2007 in the Dioila and Mande zones .............................. 104

Table 3-2: Number of farmers by gender who participated in the on-farm farmers’
sorghum evaluation exercise aimed to assess farmers’ variety preferences in the Dioila
and Mande zones in December, 2007 ........................................................... 107

Table 3-3: Farmers’ preferences expressed in percentage for the 32 varieties in the on-
farm trials by gender, variety type (short versus tall types) and by zone (Dioila and
Mande) from the ICRISAT/ [ER in December 2007 .......................................... l 10

Table 3-4: The results of reasons why farmers from Dioila and Mande zones (n=230)
rated the sorghum varieties as their three top choices from the on—farm field trials in
December 2007 .................................................................................... l 10

Table 3-5: Results of farmers’ variety traits evaluation done on December 2007 by zone
of the on-farm trials, expressed as the number of farmers and percentage of the sample
citing different reasons for their variety preferences, per zone in southern Mali ......... l 1 1

Table 3-6: Results of farmers’ variety traits evaluation done in December 2007, by
gender ................................................................................................ 1 12

Table 3-7: Results of farmers’ variety traits evaluation done in December 2007, by
gender and by zone ............................................................................... l 12

Table 3-8: The name of dominant preferred sorghum varieties (varieties grown in large
space by farmers encountered in Dioila and Mande. The table also shows the frequency

xiv

(%) (the number of time a variety was cited as dominant by farmers). The data to make
thi s was from farmers’ general survey conducted in December 2007 ...................... 1 13

Table 3—9: Results of farmers’ variety traits preference evaluation of their preferred

dominant varieties expressed in percentage, from the farmers’ general survey conducted
in December 2007 in the Dioila and Mande zones in southern Mali ........................ l 14

Table 3-10: Results of farmers’ variety traits preference evaluation of their preferred

do rninant varieties, expressed in percentage per zone from the fanners’ general survey
conducted in December 2007 in Dioila and Mande zones ................................... 1 14

Table 4-1: The main sources of sorghum seed, villages where seeds were obtained, and

the time of its acquisition by zone (Dioila and Mande), from the farmer general survey
done on December 2007 .......................................................................... 131

Table 4-2: Information on sorghum purchased seeds in Dioila and Mande, from farmer

general survey conducted on December 2007 to document market sourced seed in both
zones ................................................................................................ 132

Table 4-3: Prices (in F CFA) that farmers are willingness to pay for seeds of their three
most preferred varieties, for the entire sample (n=230) ....................................... 133

Table 4-4: Farmers’ willingness to pay for their preferred varieties survey per zone. . ..134

Table 4-5: Results of farmers’ mean willingness to pay for their preferred sorghum
varieties in relationship to the most popular variety traits studied in chapter 3. Mean price

in CF A for a kg of seed in relationship to varietal traits (the standard deviations are in
parentheses) ......................................................................................... l 35

Table 4-6: The numbers of participants interviewed at the fairs held in Bancoumana and
Siby in May 2007 in Mande zone to test the appropriateness of seed fairs as tool to
improve farmers’ access to good quality seed of their preferred varieties ................. 136

Table 4-7: Numbers of crops species presented at the seed fairs held in May 2007 in
Bancoumana and Siby in the Mande zone ...................................................... 136

Table 4-8: Information on crop types (number of varieties, variety names, variety types

and reasons of selling it) presented on the seed fairs held in Bancounama and Siby in
May 2007 ........................................................................................... 137

Table 4-9: Information collected from seed fair vendors about their experience as seed
sellers and membership in farmers’ organizations ............................................ 138

Table 4-10: Information collected from seed fairs seed buyers about their experience in
buylng seed and the reason why they chose to buy seed at the fairs ........................ I39

XV

Table 4-11: Information collected from seed fairs visitors about their experience in
vi siting the fairs held in Siby and Bancoumana on May 2007 in Mande Zone ............ 140

Table 4—12: Quantity of sorghum seed purchased at the seed fairs held in Bancoumana
and Siby in May 2007 during the time the surveys were conducted ........................ 141

Table 4-13: Buyers’ impressions on seed quality, prices and seed presentation in the seed

fa i f S .................................................................................................. I 42
Table 4-14: Suggestions made by the interviewed seed fair participant in Bancoumana

and Siby for future improvement of seed fairs in the Mande zone ......................... 143
Table 4-15 Estimate cost of production of 1 kg of improved sorghum seeds ............. 146

Table4-16: Estimated revenue for sorghum seed production and distribution per zone
when research costs are not paid for through market receipts ............................... 147

xvi

LIST OF FIGURES

.Fi gure 0-1: Map of part of Mali indicating the study zones .................................... 5
Figure 0-2: Map of Mali Republic and surrounding countries ................................. 8
Figure 1-1: Key terms definition .................................................................. 16

Figure 1-2. Farmers’ and institutional system of management of plant genetic resources
(copied from Almekinders & Louwaars, 1999) ................................................. 17

Figure 1-3: Percentage of seed impurities weight by crop for market versus farmers own
produced seeds of seed samples collected in Dioila and Mande. The bars represent the

average weight (means) and lines above the bars represent the standard deviation from
the mean ............................................................................................... 45

Figure 1-4: Percentage of germination by crop for market-sourced versus farmers’ own-
produced seeds of seed samples collected from farmers in Dioila and Mande. The bars

represent the average weight (means) and lines above the bars represent the standard
deviation from the mean ........................................................................... 46

Figure 1-5: The percentage of infected seedlings for each of three crops for market-
sourced versus farmers’ own-produced seed of seed samples collected from farmers in
Dioi la and Mande. The bars represent the average weight (means) and lines above the
bars represent the standard deviation fi'om the mean .......................................... 46

Figure4-1: Farmers willingness to pay evaluated by gender in Dioila and Mande zones

through the survey done along with on-farm variety evaluation in the two zones in
December 2007 .................................................................................... 134

xvii

KEY TO SYMBOLS OR ABBREVIATION

ACOD : Association Conseil pour le Développement (Local NGO)
AOPP : Association des Organisations Paysannes des Producteurs

A PCAM : Assemble'e Permanente des Chambre d’Agricultures Maliennes
APROF A : Agence de Promotion de Filiéres Agricoles.

AV : Association villageoise

BNDA : Banque Nationale pour le Développement Agricole

C OPROSEM : Cooperative des producteurs de semences

CFC : Common Fund for Commodities

C ILSS : Comite Inter-état de Lutte contre la Sécheresse au Sahel

C I MMYT: International Maize and Wheat Improvement Center

CM DT: Compagnie Malienne pour le Développement Textile (Cotton Company in Mali)
C IRAD: Centre International de Recherche Agronomique et de Développement
COPROSEM : Cooperative pour la Promotion de la Filiere semence

DG RC : Direction Ge'nérale de Regulation et Contréle

Diff : difference

DN S I : Direction Nationale de la Statistique et d’Information

DNAMER : Direction Nationale d’Appuis au Monde Rural,

ECOF IL : Economie des filieres Agricoles

EDS: Enquéte Démographique et de Santé

FAO: Food and Agricultural Organization of the United Nation

Fcfa : Franc Communauté Financiere d’Afi'ique

I:PVS : Farmers Participatory Variety Selection

FV 3 Farmer variety

GDP 1 Gross Domestic Product

GIE I Groupement d’Intérét Economique
GLM : Generalized Linear Model

gr 2 Gram

1 CR1 SAT : International Crops Research Institute for Semi-Arid Tropics

xviii

IDRC : International Development Research Center
113 R: lnstitutD’Economie Rural
[PR /IFRA : Institut Polytechnique Rural/Institut de Formation et de Recherche
Appliquées
IS'TA : International Seed Testing Association
Kg : Kilogram
LCV: Local control van'ety
LS M: Least Square Means
MV: Modern varieties
m2 = meter square
NGO: Non Governmental Organization
n: Total number (Sample)
OHVN : Office de la Haute Vallée du Niger
OMA : Office des Marches Agricoles
OPC : Organisation des producteurs de ce'réales.
R Var : response variable
PVP : Plant Variety Protection
PPB : Participatory Plant Breeding
ULPC : Union Locales des Producteurs de Céréales ( local farmers’ association)
SAS : Statistical Analysis Software
SSA: Sub-Saharian Africa
SSN : Service Semencier National
SD : standard deviation

SE : Standard Error
W01: Weight of Impurities

xix

INTRODUCTION

Mali, as most countries in Sub-Saharan Afi‘ica (SSA) relies on agriculture as the
backbone of its economic and social development because it employs about 80% of the
po pulation, constitutes the main source of domestic food supply and produces about 45
GDP (Coulibaly, 2004 ). Sorghum (Sorghum bicolor) and millet (Pennisetum glaucum)
constitute the main cereal crops in the Koulikoro region (the study area in Mali). They
account for more than 40% of the total food grain production in the region, with small-
holder farmers responsible for almost all the production (Tefft, 2004). In Mali, sorghum
and millet are subsistence crops for almost all of the population and the total area that
they occupy are bigger than that of all other crops. This is due to the fact that these crops

can grow well in drier regions and on poor soils than other crops (Byth, 1993).

Groundnut on the other hand, constitutes the main source of income for women
and source of food and condiments for the households in most of the study areas.
Groundnut more than sorghum and millet can grow well in dry regions and on poor soils
which are most of the time given to women (Weltzien et a1. 2006). Another important
point about groundnut is the fixation of atmospheric nitrogen which helps in restoring
soils fertility. These three crops are grown by almost all the farmers in this region.

DeSp ite their importance in Mali, all these crops share constraints related to seed quality
and access. To understand agricultural development in such an area, it is essential to look
at CT Ops that are grown by both men and women, especially the small-holders. Thus to
assess seed quality at farmer’s level in a participative way (involving all farmers: men

and Women), looking at seed of these three crops can give a good insight.

Seeds are the most important input in all crop-based agriculture and a prerequisite
for the majority of the world’s food production (Badstrue, 2007). They provide the basis
of factors influencing crops’ yield because they contain all the genetic information of
varietal traits that determine yield potential, adaptation to environmental conditions, and
resistance to pests and disease (Maredia et a1 .1999). Research has shown that healthy
seed is one of the important factors in improving agricultural production (Gupta, 1999).
Therefore, a farmer's most critical management decision is the selection of seed sources

and varieties.

In Mali as in many SSA, farmers use seeds from informal sector (about 80%)
(Siart, 2008), constituted of seeds produced by farmers without quality control and from
local varieties. These seeds, even though inexpensive, are thought by some researchers to
have some limitations, such as low germination rates, varieties with low yield and weak
resistance to climatic stress, poor pest resistance, etc... That is why some argue that the
agricultural productivity growth in SSA has not been as strong as compared to Asia,
where the development, distribution (marketing) and use of improved seeds of new
varieties were important components of the Green revolution (Maredia et a1. 1999). In
SSA, the existence of efficient seed systems including all source of seeds, e. g., a system
where all farmers can access good quality of seeds (improved seed', high-yielding,
disease-flee) of locally adapted varieties ( socio—economic and agro-ecological adapted
varieties) on time is a primary determinant of productivity (Cromwell and Tripp, 1994).

The wide use of local seeds in Mali is due in part to the fact that most farmers are

poor and therefore cannot afford new improved inputs particularly seeds

 

' Improved seed refers in this paper to seed produced under quality supervision by the national seed service

produced/developed by researcher and/or imported because of cash flow problem. They
also have limited access to agricultural services and markets because of poor
infrastructure (CIMMY T, 2007). This problem is important for grain crops produced
mostly for subsistence (Yapi et al, 2000) and for legume crops characterized by difficult
storage, disease susceptibility, low multiplication rates and high seeding rates
(Vanderhofstadt, Bruno. 2002). Thus, seed suppliers find it too risky to invest in the seed
sector because there is no sure demand (Tripp er al, 2001 ).

However, seed system analyses in Mali have shown that farmers are interested in
trying new types of varieties that combine socio-economic quality traits (storability, easy
to process, culinary quality, and marketable) (ICRISAT, 2005).

In the past, most agricultural research in Mali was focused on cash crops like
cotton and rice (Teflt, 2004). Then, with the problem of food insecurity in Mali, some
research institutions (IER and ICRISAT) started to focus on agronomic performance of
grain crops such as sorghum millet, and maize in order to increase food security for the
majority of the population. ICRI SAT has been conducting research to produce superior
sorghum germplasm and has developed some improved varieties for farmers to use.

In this new environment characterized by the existence of new improved varieties,
farmers’ willingness to try new varieties, investigating seed system functioning through
assessing seed quality and varietal preference, will contribute significantly to improving
farmers’ access to good quality seed of their preferred varieties in Mali.

This research will focus on seed quality for the following three crops: millet,

sorghum, and groundnuts. This project will investigate the status of existing seed quality

fi0m different sources for each crop by answering the following questions:

° What is the quality of sorghum, groundnut and millet seeds used by farmers?

° How is the varietal purity of improved variety seeds affected by farmers’
seed recycling?

0 Which variety traits are preferred by farmers?

° How can seed sector performance be improved, based on information
gathered through assessment of farmers’ willingness to pay for improved
seeds and their interest in seed fairs?

The answers of these questions this research will help to address some of the
major constraints to the development of a seed system that meet the needs of farmers in

Mali. These constraints include:

I) Inadequate identification of quality concerns related to the recycling of seeds
over several years, including the conditions of production and storage of seeds that most

farmers use in Mali for sorghum, groundnut, and millet,

2) Difficulties in planning and programming seed production in relationship to the

users’ needs and varietal preferences, and

3) Inadequate knowledge and limited dissemination of information about
varieties, particularly for these subsistence crops and “women’s crops” (such as

groundnut).

This project will contribute to the broader question of how to develop sustainable
seed systems for low-income tropical countries in general. It will help African policy

makers, NGOs, and donors to identify:

0 Strengths and weaknesses of the current seed systems for these crops

o The critical areas for further investment and training
This research was conducted in two zones: Mande and Dioila in the Koulikoro
region in Southern Mali (see figure 0-1). In each zone, four villages were selected and
within each village, 20 households were selected. From each of the participating
household seed samples were collected for field trial and laboratory analyses. The head of
each household was interviewed as well as each individual grower of sorghum, millet and
groundnut in the household. The selection methods for the villages and households are

described in the next chapter (chapter 1)

... ,/ " '-\,
M a n d .mpmmoo «womb-clan ‘ ”mm )
It." , >rnd'AIm-n 3*. ram»; -~- _";7' .. G A 0
. «‘ T ' . - ’ am". . I .le
__ . 1‘ g . \ I
i " J‘./>-\f"-l\ \ V-i
" ‘ , “WV I’ Am: 5 , 1’" \ \W!“ _
1‘ m j ' K ‘ IOPII .m .r r.
“a“ ‘ ‘ O ”“7" '5 “W 3...... ,. ~. N I G’s R ‘
' R mvts 3» . . «I... 1_ 711.». . - ,3. "A; ,‘
“A —"‘-“~ ‘ KK'JIJJI‘I'IWIL. SEOUL A:/' - \ ~ '1' {V . . \k ‘-\‘,. 1
16m I \x ' ‘ 11.?th " . v‘y-x , ”'7" M,
m'.<-. wlm' .3,‘ _> ,- B R K‘I‘ N A $
.. \y ,1 \*/\fi i/‘flif‘ i‘m', U ‘ I k ' __
‘l . . '-. SEW j. N». I, l
(i 5 ’ fl. PM)" ’ ' “ 17‘7"? ”-4 ‘7‘; "\
\rl L ' J» w; ‘ -‘ ,.--j_’.» ' X'. I possum-k . '
s 'x. - ‘./*»~~s\;;'13 r ~_‘$‘~“r‘ 8“,. r: _/ h“ ‘ I I! . ‘__\ 1.' -.
GflNEA , , wasso “‘f , ’_ u,“
:5"; ._ sw ’_ , .,,,_,T.. 9- , '. \ __ ‘1
\V‘. ‘ I .‘_ Gnu; . '3 ' ‘I‘ 4km ‘7
m..3.>\ . 5 U ‘ A 1‘ .i . ‘ “\B E N ‘ N v":
D'o"|

Figure 0-1: Map of part of Mali indicating the study zones.

Source: Sonja Siart, 2008

The data (quantitative and qualitative) used to answer these questions were
collected through both formal and informal surveys at farmers’ level, field trials, and lab
qu ality test on seeds collected from farmers, researchers, and local seed producers. These
data were used to assess seed quality from different sources (own production versus

market or research), different storage methods (grains versus panicles/pods), different

locations (Dioila and Mande), across the three crop types, and farmers’ varietal
preferences. I also used data collected during the seed fairs and variety evaluations to

assess the possibilities to improve farmers’ access to good quality seeds.

This dissertation is consists of four chapters in addition to this introductory part.
Chapter 1 deals with farmers’ seed quality documentation, and chapter 2 focuses on
farmers’ seed recycling for four ICRISAT improved varieties in production at farmers’
level since 2003. Chapter 3 assesses farmers’ preference for different variety traits among
the improved sorghum varieties in on—farm-field trials. And chapter 4 examines ways to
improve farmers’ access to good quality seed of their preferred varieties through seed

fairs and farmers’ willingness to pay for varieties with their preferred traits.

OVERVIEW of SORGHUM, PEARL MILLET and GROUNDNUT
Sorghum and Pearl millet

Sorghum (Sorghum bicolor) is a diploid (2n= 20) and largely self-pollinated crop
(2-20% outcrossing) (Minot et al., 2006). Sorghum is a 04 species adapted to dry
conditions, with minimum requirement of about 500 mm of rainfall (Minot et al., 2006).
According to the same source, sorghum is grown on 26 million ha in sub-Saharan Afiica,
with an average yields of only 800kg/ha due to low harvest index (FAOSAT, 2007). In
West Africa, sorghum is an important dietary staple crop. The stover is valued as animal
feed, fencing material and firel in mixing cropping systems (Folkertsma et al., 2005).

Pearl millet (Pennisetum glaucum) is the most important of the millets and is the
preferred cereal crop in the sahelian region of West Africa (Minot et a1., 2006). Millet is a
diploid species with 2n=14, and is an open-pollinated crop with 75% to 80% outcrossing.

Like sorghum, millet is c4 species adapted to dry conditions. The early-maturing varieties

have a growing cycle of about 60 days, while the late-maturing ones can go up to 180
days. Millet is native to Afiica, and West Afiica is its center of diversity (Minot et al.,

2006).

Groundnut

Cultivated groundnut (A rachis hypogaea L.) belongs to genus Arachis in sub-tribe
Stylosanthinae of the tribe Aeschynomenea of family Leguminosae. It is a self-pollinated,
tropical annual legume (Nigam et al., 2004). Groundnut is the sixth most important
oilseed crop in the world. It contains 48- 50% oil and 26-28% protein and is a rich source
of dietary fiber, minerals and vitamins (N ’Tare et al, 2008). The optimum temperatures
for growing groundnut range from 25°C to 35°C. Early-maturity varieties require 300-
500 mm, while late-maturing ones need 1000—1200 mm rainfall. Groundnut produces
well on well-drained sandy loamy soils.

Groundnut is grown on about 26 million ha worldwide, with a total production of
about 37 million metric tons and an average productivity of 1 t/ha (FAO, 2003).
Groundnut is grown in over 100 countries worldwide, with 97% of the global area and
94% of the global production in developing countries. The production is concentrated in
Asia and Afi'ica, with 56% and 40% of global area and 68% and 25 % of the global
production respectively Wtare et al, 2008). Grain legumes including groundnut are
widely grown by smallholder farmers in many semi-arid countries like Mali because they
derive multiple benefits from growing legumes grains, including nutritious food, soil

fertility and cash income.

 

STUDY AREAS

The Republic of Mali is a landlocked country located in West Africa. It is
surrendered by Algeria, Burkina Faso, Guinea, Cote d'Ivoire, Mauritania, Niger, and
Senegal. A former French colony, Mali gained independence on September 22, 1960,
with Bamako as its capital. Mali is composed of eight administrative regions: Gao,
Kayes, Kidal, Koulikoro, Mopti, Segou, Sikasso, and Tombouctou. With a population
estimated to be about 12,000,000 people (EDS/DNSI, 2004), Mali has a total area of 1.24
million sq km, of which 1.22 million sq km is land and 20,000 sq km is water (DNSI,

2005)

 

 

 

 

 

Figure 0-2: Map of Mali Republic and surrounding countries
Source : htti:,//www.cia. rov/cizu’ ublicatioIIs/factbook/

Icos/mllitml

     

The climate is subtropical to arid; hot and dry from March to April; rainy, humid,
and mild May to October; cool and dry November to February. The Niger River cuts an
arc through the country.

Mali is among the poorest countries in the world, with 65% of its land area desert
or semi-desert. Economic activity is largely dominated by the agricultural sector, which
contributes about 45% to the GDP and ensures work for about 80% of the active
population.

Malian agriculture has been focused for a long time on grain crops (millet,
sorghum, maize,. . .) and cash crops (cotton, peanuts,. . .), as well as ruminants. The type
of crops grown in each region depends on traditional cultivars.

ICRISAT Research Station in Samako (Mali)

The ICRISAT -—Mali center is situated at Samako, which is 25 km from Bamako.
It covers a total land area of 124 ha, much of which is set aside for on—station variety
trials. The latitude and longitude are 12° 54” North and 8° 4” West respectively. The
average altitude is about 328 to 330 m above sea level. The climate is typically Sudanean,
with a harsh dry period stretching fiom March to May, and a rainy season from June to
September.

One of ICRISAT’s objectives is to improve productivity and stability of
agricultural production of grain crops such as sorghum, groundnut and millet by
providing access to new varieties and germplasm (ICRISAT 2002 in Siart et al., 2005).

Participatory plant breeding2 is one of the effective ways that ICRI SAT is using to

 

2 Participatory plant breeding is a research method in which scientists collaborate with local farmers in
order to share their knowledge and to find out how ICRISAT can optimally meet their needs.

achieve this objective. This research uses ICRI SAT’s participatory research method to

assess seed quality with farmers in two of its intervention areas: Diola and Mande.

Study zones

The research was conducted from May 2007 to September 2008 in the sorghum-
based production systems of southern Mali (Siart, 2008). Southern Mali accounts for
more than half of the millet and sorghum supply in Mali (ABT, 2002). The research was
conducted in two zones of this part of Mali:

0 The Dioila zone, located 160 km southeast of the capital Bamako, is a Malian Cotton
Company (CMDT) zone.
0 The Mande zone, about 60 km west of Bamako, is an “Operation Haute Vallee du
Niger” (OHVN) zone
The two zones are similar agro-ecologically, but differ widely in the degree of
mechanization of agriculture, market orientation, infi'astructure and organizational
environment (ICRISAT, 2002) as shown in table 0-1. Dioila is considered as the more
developed region because of the presence of CMDT, the well structured and oldest cotton
company in Mali.

The average annual rainfall is about 800- 1000 mm, and the length of the rainy
season varies between 4 to 5 months, starting fi'om May or June to September or October.
As in most Sahelian countries, the starting of the rains is very uncertain in the study
zones. Throughout the country, from south to north, the irregularity of the rainfall
increases as well as the duration of rains, and the overall quantity of rain is decreasing
(SIMPSON, 1999). This situation has contributed to marked changes observed over the

recent decades in the agro-ecological conditions in the growing areas for rain-fed crops

10

like sorghum, millet and groundnut, including the study zones. Since the severe droughts

of the early 19703, the average rainfall in Sahel in the 30 years up to 1997 was 20- 49%

lower than in the period between 1931 and 1960 (IPCC, 2001 cited by Siart, 2008). Soil

fertility is also changing due to shorter fallow periods and insufficient use of fertilizers on

sorghum, millet and groundnut fields. (Weltzien et a1. 2006)

Table 0-1: Socio-Agro-Ecological conditions of Dioila vs. Mande

 

 

Conditions Dioila Mande
Distance to Bamako 160 km southeast of Bamako 60 km west of Bamako
Major extension CMDT: the Malian Cotton OHVN: Upper Niger River
service Company Valley Development

Characteristic of
agriculture

ICRI SAT
partnership

Ethnicities
Agro-ecological
characteristics

More Mechanized

More market oriented
More organized farmers
More infrastructure (road,
storage places, etc...)

ULPC : Local union of coarse
grain producers

Mainly inhabited by Bambara
800 — 1000mm of rain/year, with
a decreasing length of rainy
season in the recent decades.
Soil fertility is also decreasing.

Authority
0 Less Mechanized
0 Less market oriented
farming
0 Farmers are less
organized
0 Less infi'astructure

ACOD: National NGO
AOPP: Association of
professional organization of
producers

Mainly inhabited by Malinke
Same as in Dioila, but
sometimes more rains.

Soil fertility conditions are
similar but sometimes worse
in Mande because farmers do
not have access to fertilizers as
they do in Dioila

 

Brief Overview of Agriculture in the Study Zones

Both areas are cotton-growing zones, but Dioila is one of the oldest CMDT

regions in Mali. In Dioila Zone, farmers have access to inputs (cotton seeds, fertilizers

and pesticides) and basic farming equipment (ox plows, planters, etc.) on credit for cotton

11

production. In Mande, on the other hand, the cotton production is not as important as in
Dioila. It is under OHVN extension service, and farmers do not have access to credit for
cotton production. As result, the level of mechanization in Mande is lower compared to
Dioila (Siart, 2008).

Sorghum is the main staple crop in both zones, but other cereals like maize,
millet, and rice are also grown. A typical cotton-growing household’s cropping system is
as follows: cotton—maize—sorghum—millet—groundnut (grown by women). If the
household doesn’t produce cotton, maize takes the head of the system (Weltzien et a1.
2006). In this cropping system, sorghum, millet and groundnut are grown on poor soils
that are not suitable for cotton and corn. However, if grown in rotation with cotton,
sorghum can profit fiom the residual fertilizer.

In both areas, farmers grow sorghum, millet and groundnut in two types of fields:
bush field and house field. The bush fields are usually far from the village. Due to
transportation problems associated with long distances, these fields usually receive only a
small portion of the households’ manure. The house fields are located near or within the
village and receive more manure because they constitute the place where animals are
tied-up at night during the dry season. They also receive more attention because of their

location. (Weltzien et a1. 2006).

Socio-Economic and Cultural Situation of the Research Zones

Sorghum and millet are grown mainly for subsistence, more specifically in Mandé
and with only some degree of commercialization in the Dioila region (Siart, 2008). The
cereal producers’ organization in Dioila (OPC) coordinated by ULPC (Local Union of

Cereal Producers) makes this market orientation for sorghum possible in this region.

12

Cotton is the main cash, crop followed by groundnut for men; groundnut is the main cash
crop for women in both zones.

Sorghum and millet are primarily men’s crops and are grown in family fields
where all able-bodied family members (expect the old women) are required to work
(Weltzien et a1. 2006). The lack of labor is an important limiting factor to agricultural
production in both regions because more people tend to do more off-farm income-
generating activities from which they can get more income than from agriculture.
Children and young people attending school, which might be outside the village, also
reduce the amount of labor allocated to agriculture (Weltzien et a1. 2006).

The two zones are different in terms of the presence of organizational structures.
Dioila has more farmers’ associations than Mande', which has practically none. This
difference is due the fact some external organizations working in Dioila like CMDT and
SNV have identified farmers associations as an important agricultural development
factor. As a result, they helped in initiating them in their intervention areas (villages)
(Siart 2008). The Mandé region, located on the upper Niger River and the border of
Guinea, is seen as a heartland of the old empire of Mali (the empire led by Soundiata
Keita). This place still plays a big role in the history of the country. The Mandé language
group accounts for about 40 languages in West Africa (Mabe, 2001 cited by Siart 2008).

The farm households are usually extended, large families headed by one head of
family with his sons’ and brothers’ families. All the active family members work in the
fields where sorghum, millet, and maize are grown for family food and cotton for cash to

support family expenses (Weltzien et a1. 2006).

13

The traditional structure of the society is extremely hierarchical. At the village
level, the head of the village (chef de village) is from the founder family. No individual
land ownership exists; all lands belong to the community, and only the land use right is

inherited (Linding, 1986 cited by Siart 2008).

’14

CHAPTER 1

QUALITY of FARMER PRODUCED SEEDS of SORGHUM,
GROUNDNUT, AND PEARL MILLET IN DIOILA AND
MANDE ZONES

INTRODUCTION

Mali’s economy, like that of most SSA countries, is largely dominated by the
agricultural sector, which contributes about 45 % to the GDP and ensures work for about
80% of the active population. Agriculture constitutes the main source of domestic food
supply and ensures at least one third of foreign exchange earnings (Diakite and Diarra,
2000). Malian agriculture is representative of the path of development for SSA countries,
where subsistence, extensive cropping systems still are dominant, with minimal use of
improved inputs including good quality seeds (David, S. and L. Sperling. 1999). In SSA,
the existence of efficient seed systems—Le, systems where all farmers can access to
good quality of seeds (improved seed, high-yielding, diseases free) of locally adapted
varieties on time— is a primary determinant of productivity (Cromwell et al., 1992 and

Tripp 2000).

15

Seed: Parts of agricultural, silvicultural, and horticultural plants used for sowing
or planting purpose

The four basic seed quality aspects are:
1- Physiological quality (germination, vigor)
2- Sanitary quality (absence of seed-borne diseases)
3- Analytical quality (percentage of good seed in a particular seed lot
4- Genetic quality (varietal adaptation, varietal purity

Good quality seed: refers to seed with good germination rates, seedling vigor,
absence of weed seeds or seed-borne diseases and other impurities.

Seed system: the entire complex of organizations, institutions, and individuals
associated with seed program of a country comprised of the traditional or
informal system of farmer-selected, —multip1ied, -processed, exchanged and
retained seeds, and a non-traditional or formal system of individuals,
organizations and institutions involved in specialized tasks related to producing
and marketing seed for sale to seed users.

Figurel-l: Definition of Key Terms

Farmers’ seed sourcing in Mali

Mali as in many SSA countries, farmers get their seeds from mainly two sources:

informal/local sector and formal sector as described in the Figure 1-2

16

 

Inspections

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Gene Bank \
\ Logistics,
chain
Phytosanitary management
inspection and —' Breeding and storage
certification
. 4 T
u. Multiplication : l
O. .0
O O
..O 0.. '
O O
.0. .9. I
’0“ Certification ..‘° I
O . . ’ . O
0......-.-"-....‘ I
. /
Farmers (clrents of the W
produced seeds)

 

 

 

— '> Flow of certified seeds to farmers

_, Flow of seeds from production to storage or logistics
—> Inspections of seed production process

- - - - - I) Seed certification

Figure 1-2: Formal seed system description adapted from Willy De Greef Biotechnology
Specialist, PRODEPAM project, 2004

Formal seed system in Mali

The formal seed system is characterized by vertically organized production,
storage and distribution of tested seed and approved varieties, using strict quality control
(Maredia et a1., 1999). In Mali, the most important formal seed channel for millet and
sorghum and groundnut is the National Seed Service (SSN). SSN originated as part of the
state-owned production and distribution system for seed (Minot and Smale, 2006).
Today, the formal seed system is coordinated by the SSN, which has public and private
components. The public components include governmental structures of research and
variety development such as IER (Institut d’Economie Rural), extension (DNAMER—
Direction National d’Appuis au Monde Rural), quality control (DGRC—Direction
Generale de Regulation et Controle), and the SSN (Service Semencier National). The
private components include private seed farms and seed producer farmers’ groups
(Diakite, 2003).

The formal seed system in Mali has focused on cotton, rice and maize. To a lesser
extent, work has been conducted on developing a formal seed system for millet, sorghum,
cowpea and groundnut (Tefft, J. 2004). The seed quality requirements recommended for
these crops are to obtain >80% germination and >90% physical purity (Dembele, 2006).
The use of formal system seeds is still almost nonexistent for millet, groundnut, and
sorghum, which is related to many of the reasons discussed in this research. An earlier
study (Yapi et a1, 2000) indicated that a formal seed system is not developed because of
the high cost of certification, which limits seed affordability for resource limited farmers,

and because farmers can source seeds from the informal system (Maredia et a1., 1999).

18

Informal seed sector in Mali

The informal seed system consists of farmer-to-farrner seed exchanges and
farmers’ own saved seeds produced from local varieties without any supervision or
quality control during production (Pejuan, 2005). This system is based mostly on
farmers’ saved seed from their harvests, but local markets may play an important role for
seed security as well. Seed and consumption grains are grown in the same field, but
some farmers select and store seeds separately from the gain (Minot and Smale, 2006;
Bazile, 2006; and Coulibaly, 2004).

In Mali, as in many Sub-Saharan African (SSA) countries, farmers obtain almost
all their seeds through the informal sector (Siart, 2008, Almekinders and de Boef, 1999).
Getting information on how small-scale farmers obtain, manage and share seed is crucial
for designing appropriate mechanisms for delivering new crop varieties (David, 2003).
Assessing farmers’ seed quality for a range of crops (sorghum, millet, and groundnut)

will help in documenting the seed quality in the study areas.

Farmers’ local seed management3 for sorghum, groundnut and millet in southern
Mali

Seed production

For the most part, farmers select their own seeds in Mali. A sorghum study done
by Siart (2008) in the two zones of Mali showed that 82% of sorghum seeds were
selected by farmers in the field as panicles before the harvest (in Mande). In Dioila about
50% of seeds are selected after harvest. These results are similar to findings by Diakité

(2003), although in Dioila the number of farmers selecting seed in the field has slightly

 

3 Partner seed management refers to the process that farmers use, to produce. to obtain. to maintain. to
develop and distribute seed resources.

19

increased. Diakité (2003) reported that three-quarters of the farmers in Dioila were not

doing seed selection in the field.

For millet, seed production at the farmers’ level has not been documented in Mali,
but from what we heard from farmers in a preliminary survey to this dissertation
research, selection of seed usually does not occur in the field. The seed is selected from
stored grain, at planting time. Only a few farmers reported selecting the best panicles of

millet during harvest to save as seeds for the next planting season.

In contrast to sorghum and millet, for groundnut there is no seed selection in the

field.

Seed storage

Sorghum seeds are stored by smallholder farmers in SSA in various ways, such as
grains in bags, grain in granary, panicles attached together (sheaves) stored in the kitchen
or outside, and inside the granary. Often panicles are stored in bags in the granary. These
are the storage systems used in southern Mali as well. In Mande, the majority of sorghum
seeds are stored in sheaves (71%), and about 19% are stored as grain. In Dioila, on the
other hand, only 30% of sorghum seeds are stored in sheaves and 25% are stored in
panicles not attached together either in bags or granary (Siart, 2008). Millet seed storage
is quite similar to sorghum in both zones. The only difference is that millet is mostly
stored in grain. Groundnut seeds are stored in pods (the majority) or as grain. Farmers
usually store only limited quantities of groundnut in pods to use as seed because of

financial and logistical reasons (Ntare et al, 2008).

20

Groundnut seed management by women

Groundnut constitutes the main crop for women in both zones (Brock, 2007).
During the surveys, the informal discussions with old women growing groundnut
revealed that almost all of them have their dominant groundnut variety since their
wedding. According to them, “in our culture when a girl gets married, her mother should
give her some groundnut seeds to start her field. It is her duty to multiply and maintain
the seeds and provide her own daughter with seeds as wedding gift.” These women have
kept those seeds for more than 40 years (author’s unpublished survey data, 2007) and
they were still visually looking good. In the study areas, men’s groundnut seeds are
usually from women’s production even if seeds are purchased (informal discussion with
Awa Traoré of ICRISAT Mali). Thus, all these observations suggest that for a crop in
which production both women and men are involved, like groundnut, women’s seed

quality is likely to be better compared to men’s.

Biological differences between cereal crops and oilseeds

An important difference between the cereal crops sorghum and millet and the
oilseed crop groundnut is that the amount of seed required to plant the former is much
smaller than the later. The quantity and cost per hectare is quite small for cereals— about
4—8 kg of seed for sorghum and 3 -6 kg of seed for millet. Thus, for cereals the quantity
of seed that farmers save is usually sufficient to plant their fields (Minot and Smale,
2006). It is also possible to obtain sufficient seed to plant through gifts from kin or
friends, since the quantity is not large. That is one reason why seed purchases are rare for
these cereals and the commercial formal seed sector has not developed for sorghum and

millet in Mali, and indeed in SSA generally (Minot and Smale, 2006, Tripp, 2000).

21

The biology of groundnut necessitates that seed inputs are a substantial
investment in production by farmers. The species is characterized by large grain size, and
high oil content, increasing vulnerability to degradation during storage as well as
markedly increasing the expense associated with groundnut production. The quantity of
seed needed to plant one hectare is about 80-100kg/ha depending on row spacing, plant
spacing, seed-mass, and percentage of germination of the seed sample used (Singh and

Oswalt 1995).

Seed quality parameter assessed in this chapter

Cushman (2006) defined quality seed as varietally pure seed with a high
germination percentage, free from disease and disease related organisms and with a
proper moisture content and weight. Seed quality is determined by many factors,
principally seed purity (physical and genetic) and physiology (level of maturity,
capability to germinate) (Brick, 2004). However, many other factors, such as the presence
of seed-borne disease, seed conditioning (storage, age, packaging), and size influence
quality (Copeland et a1. 2001). Spreading out a representative sample in a single layer on
a table top provides an excellent opportunity to observe physical quality and condition of
the seed. Laboratory analysis, however, is the only reliable means of determining the

ability of seed to germinate, and the presence of diseases (Basra, 1995).

Seed physical purity is determined by the amount of unwanted material present in
the pure seed. Seed should not be mixed with other classes or crops, especially those that

mature at the same time. The best quality seed is nearly 100 percent pure (Brick, 2004).

22

Seed germination tests: germination is in another word the sprouting of a
seedling from a seed of an angiosperm or gymnosperrn. The test of germination assesses
seed ability to produce new plants when placed under favorable conditions of adequate

moisture, temperature, and oxygen (Copeland et a1. 2001).

Seed-born diseases or health status tests assess the ability of the seed to produce

a healthy plant when placed under favorable environmental conditions (Basra, 1995).

There are other factors to test seed quality such as: test weight per 100 grains,
kernel plumpness, color, and conditioning (storage and handling) (Brick, 2004). For this
research I will focus on the first three factors to assess farmers’ seed quality, as shown in

table 1-1.

Table 1-1: Seed quality parameter characteristics based on seed certification standards
used by the International Seed Testing Association (ISTA) and the “Comite Inter-état de
Lutte contre la Sécheresse au Sahel” (CILSS)

 

 

Parameters Physical Physiological quality Health Need of
purity status seeds /ha
Crops Weight of Germination Weight of Infection to plant 1 ha
impurities 100 grains
(%) (%) (g) (%) (kg/ha)
Sorghum <2 80 2.8 <2 :8
Millet <2 80 1.2 <2 :3
Groundnut 4 70 24 2 =1 00

 

Source: N’tare et al., 2008 and Diakité 2003

23

Research Objectives

The overall objective of this research is to access seed quality of sorghum, pearl
millet and groundnut used by farmers in two sorghum production zone in Mali in order to
assess whether the quality of farmer produced seed is a factor contributing to low yield

levels in Mali.
Specific objectives:

0 Determine if high quality sorghum seed is available from the informal sector,

markets and farmer production
0 Evaluate seed quality traits in modem versus local varieties of sorghum

0 Determine which location, Dioila versus Mandé, produces better seed quality for

each crop.

0 Compare farmer storage methods in relationship to seed quality, for each crop

Determine which seed quality factors are influenced by crop species

Research Hypotheses

1)- Market-sourced seeds are better quality, i.e., having fewer impurities, less

disease infection and higher germination rates compared to farmer-produced seeds

2)- Seed quality of modern sorghum varieties is better than that observed in local

varieties of sorghum

3-) Seed produced in the Dioila region will be higher quality than seeds produced
in Mandé, as farmers have greater access to information sources in Dioila from the long-

term presence of extension, NGOs, and the CMDT (cotton company)

24

4-) Farmer seed stored in panicles or pods will be higher quality than seed stored

in grain storage containers

5-)Groundnut will be associated with the poorest seed quality, due to the high oil
content of this species, which increases vulnerability to disease and insect damage

compared to sorghum and millet seed

6-) Women will do a better job in producing good quality groundnut seed than

men, as this crop is traditionally a crop produced by women

25

MATERIAL AND METHODS

This research falls into the category of baseline studies and systems analytic
description models. It will use a step-by-step approach to assess farmers’ seed quality for
sorghum, millet and groundnut. It will then pull together what is known about the
parameters of seed quality for the three crops in Dioila and Mande' zones while
identifying the problems and good aspects of farmers’ seeds for these crops in Mali. To
do so, it uses a descriptive approach to first assess the quality of existing seeds that
farmers are planting and a comparative approach to compare seed from the two zones and
different storage methods through lab-analyses of seed samples collected at farmers’
level. To complete, it assesses farmers’ own perceptions about seed quality, seed
production and storage through two level surveys (head of households, and individual

growers) and follow-up workshops.

Choice of the research areas

The fieldwork was done from May to December 2007 in eight villages and at the
ICRISAT research station in Samako in the Koulikoro region in Mali. The laboratory
analyses on seed samples were conducted at the research station. The seed sample
collection and the surveys were undertaken in two zones: Dioila and Mandé in the
Koulikoro region.

Dioila and Mandé were chosen to conduct this research. The reason for this
choice is that ICRISAT’s local partners in these two zones are also located in these
villages. ICRISAT and [ER have been working with partners in these two areas since
1999 (Siby) and 2001 (Dioila) on seed system issues. In both areas, this collaboration has

led to the identification of superior varieties by farmers. Subsequently these varieties

26

have been inscribed in the National variety catalogue of Mali, and farmers’ organizations
have started the production of certified seeds of these varieties for marketing. These areas
thus provide an ideal opportunity to test the mentioned hypotheses. The collaboration of
ICRISAT with these organizations makes it easier to reach farmers because they are used
to the surveys and others related research activities. In Mandé, we worked with ACOD,
Office du Haute Vallée du Niger (OHVN) and AOPP and in Dioila with ULPC.

Four villages per zone; Keniero, Siby, Gonsolo and Siranikoro in Mandé and
Seribila, Mangnambougou, Wakoro, and Wobougou in Dioila were chosen. The choice
of these villages was based on the size, .i.e., the possibility to get 20 households in a
village, and the accessibility of the village during the rainy season. They are also villages
where ICRISAT and [ER are conducting some field trials with some farmers involved in
the participatory variety improvement program. Therefore, farmers in these villages are
more familiar with research works, such as interviews. The choice was made with
Moussa Kanoute (ICRISAT technician) and Dr. Eva Weltzien Rattunde, who know these

two zones very well.

Field Activities
The field activities for this part of the research included:
0 Seed sample collection
0 Lab analysis on collected seed samples from farmers in the research areas.
0 A Survey at the farmers’ level including individual surveys of sorghum ,
groundnut, and millet producers and a survey of the heads of the household from

which seeds sample were collected.

The data used to test the hypotheses were from the lab analyses of the seed
sample collected and surveys at farmers’ level. Tablel-2 gives the detail of the place

where the different activities were conducted.

27

Table 1-2: Field activities conducted in the two zones: Dioila and Mande where the
research was conducted from May to December 2007

 

 

Places Activities Approximate
dates
Séribila Seed collection End of May 2007
Dioila Mangnambougou Surveys (head of November 29 to
zone Wakoro households, and December 10, 2007
Wobougou individual survey)
Gonsolo Seed collection Beginning of June
Mandé Keniero Surveys (head of 2007
zone Siranikoro households, and November 18 -27,
Siby individual survey) 2007
ICRISAT research station Seed quality lab- October to
(laboratory) work December 2007
Dioila Follow-up September 11- 13,
workshops 2008
Siby Follow-up September 17-1 9,
workshops 2008

 

Data Collection Methods Used

The data used in this research include both primary and secondary data on seed
quality of sorghum, millet, and groundnut. The primary data are cross-sectional, both
quantitative and qualitative, which were collected through fieldwork. Questionnaires and
discussion guides were made for each group of participants according to the type of
information needed at a given level of seed production and utilization. The secondary
data are based on the results of previous research or studies on seeds in Mali or elsewhere

and some secondary data collected at the research station.

Seed Sample Collection

Seed samples for lab analyses were collected from farmers participating in the
surveys from the two zones (Dioila and Mandé). In each zone seeds were collected in

four villages and from 20 households in each village. The same villages and households

28

were used for the field survey because I wanted to get information on seed quality from
the same people to make the comparison between the information from lab analysis and
farmers’ points of view on seed quality.

The seed samples included all the three crop types: sorghum, millet and
groundnut. In each selected household, seeds for each of the crop were collected. The
head of the selected household was asked about the type of crops grown in his/her
household during the previous year and those that were to be planted during the current
planting season. The sampling was conducted during the period just preceding sowing,
and the beginning of sowing in the two target areas. Farmers were asked to share with us
samples of the seed that they were going to use for sowing in the near future. According
to his/her response, the samples were collected. This way we got, for example, two
sorghum samples, four groundnut samples and one millet sample from some households
and in others only sorghum samples were obtained.

Appendix 1 present the number of samples collected per zone and per village for

sorghum, millet, and groundnut.

Type of seeds (samples of seeds to be collected)

The seeds were collected to represent the seed being used for sowing by the
farmers in these two typical regions of sorghum cultivation. In addition to farmers’ own
seeds, we included samples of certified seed produced by farmer seed producers and seed
from crop breeders, who provide seed producers with foundation seeds.

The seeds collected represented: a) Traditional seed system seeds composed of
farmers’ own seeds, i.e., part of their previous harvest saved to be used as seeds, farmer-

to-farmer seeds, which are seeds obtained from exchanges, as gifts, as loans and/or sales,

29

seeds from local markets grain sellers; b) formal-sector seeds, including improved seeds
of improved varieties from researches, seed companies, para-statal seed service

(governmental led seed systems), and seed from farmer seed producer cooperatives.

Lab Analyses of Seed Quality Parameters

Seed samples were assessed for the three following quality traits using lab analyses to
test:
1- Physiological quality (germination, vigor)
2- Sanitary quality (absence of seed-borne diseases)
3- Physical purity (presence of good seed in a sample, no impurities such as weed
seeds, pieces of wood, etc.).
There were pictures of sorghum grain infected by anthracnose; and the signs of
anthracnose observed by microscope of the seed sample were assessed based of the visual
appreciation of these pictures. These pictures are a reference used by ICRISAT Mali to

assess anthracnose infection in sorghum seed samples before planting.

30

Table 1—3: The seed quality measurement parameters used for the laboratory analyses
conducted over the period October-December 2007 on seed samples collected from

farmers in Dioila and Mande, Mali in May 2007.

 

Parameters

Components

Measurement activities

 

Physical
parameters

Physiological
parameters

Phyto-
sanitary
parameters

Physical purity: near absence
of inert matters (stones,
sand), absence of noxious
weed seeds, and broken
seeds.

Good germination for the lab
work samples

Absence of visible fungi,
other disease and living
insect

Absence of infested young
plant after germination

Manually sort each seed sample to
separate the inert matters, weed
seeds, and broken seeds. Then
weight each lot a part and express
this weight in term of % of the
total weight of the sample

Take 100 grains from the a
physically good seed lot, put 50
grains in a petri-dish with paper
filter in the bottom for sorghum
and millet and 20 grains/petri-dish
for groundnut. Let them
germinate for 3 days for sorghum
and millet and 5 to 6 days for
groundnut then count the young
plants and express them in % by
adding the number per petri-dish

Each sample of sorghum was
observed with microscope for
presence of anthracnose which is
a fungal disease encountered
commonly in sorghum in Mali
After germination the young
plants presenting the sign of
fungal infection (covered with a
white layer) were counted for
each petri-dish

 

31

Surveys

Structured surveys were conducted in eight villages (four in each in zone). These
surveys included both open-ended and closed questions. My goal was to assess seed
quality in a participative way; thus open-ended questions were designed to allow
respondents to answer in their own words concerning what they saw as essential aspects
about seed quality, from production to storage to utilization in planting.

The survey is composed of two parts: a head-of-household survey (survey
instrument is in appendix 4), and a survey of individual sorghum, groundnut and millet
growers if they were different from the head of the household (appendix 5).

The first part of the survey was conducted with heads of households to evaluate
socio-economic information about the household: size of the farm, number of people
living in the household, number of workers in the farm, farm equipment and family
member’s participation in local farmers’ associations or organizations. We also asked
about the number of people in the household individually growing groundnut, millet,
and/or sorghum. This survey built a relationship with the household head and enhanced
farmer interest in participating in the seed survey by reading the consent statement and
asking for participant agreement first before conducting the interview. This process was
somewhat new for the farmers and helped reassure them that we were interested in their
participation and learning from them.

For this first part, the target was the 160 households from which we collected the
seed samples; we were able to survey154 of the original 160 households from which we
obtained seed. Between the seed sample collection and the survey, a few people moved to

another place, and one head of household passed away, but we were able to revisit and

32

survey from a high percentage of the households surveyed. See appendix 2 for a
description of household locations and gender of participating farmers.

A second survey was conducted with the individual sorghum, groundnut, and
pearl millet growers from each household. This survey was composed of five parts
addressing these following topics: seed source, seed production (if own production), seed
quality parameter identification by farmers, farmers’ preferred varieties, and general
information about the farmer and his/her field. If seeds were sourced from the market, we
asked questions to elicit information on this farmer choice, and to find out where the
market was located. The final survey section addressed seed quality assessment from the
farmer’s point of view, which was assessed twice - before planting and after planting.

The surveys in Mande were done from November 18 to 27th 2007 and the ones for

Dioila were conducted form November 29th to December 10th 2007.

Choice of the household participants

The choice of the household was made with the extension agent of the villages.
Together with the agent, we went from house to house to explain the objective and
activities of the research. The choice of the household was then made based on its
willingness to participate in research activities. So the households in the samples were
composed of farmers conducting on-farm field trials with lCRISAT/IER as well as
farmers not involved in the research activities but willing to participate in this research.
In all selected villages the majority of households agree to participate in the research, 20
households were randomly chosen among them. The choice of 20 households was based

on the resources (times, money, and staff) available.

33

I used this sample method based on the agents’ advice regarding previous
experience of research activities conducted in the villages—i.e., they suggested
households that had a good record of participating in previous research studies. The
extension agents did not want me to end up with a lot of household dropping out of the
study in the middle of the research.

Although the sample was from the majority of the households in the villages, the
results need to be taken with caution because we don’t have the opinion of the minority
households who did not want to participate in the research. They might have different

point of view on seed quality than the majority.

F allow-up workshops

To get the participant farmers’ reactions to the results on the lab analyses of seed
samples collected in 2007, on September 11 -19, 2008, follow-up workshops were
conducted in two places: September 11 -l 3, 2008 in Dioila in Dioila zone and September
17-19, 2008 Siby in Mande zone. In each place, two workshops were conducted where
farmers from two villages participated. In each village, 20 people (10 men and 10
women) from the 20 households from which the seed samples were collected were
invited to attend the workshops.

Each workshop lasted two days. The first days, the preliminary results were
presented and all the participants were allowed to comment on the results. The second
day, the participants were divided into four groups. Each group was given questions to
discuss for one hour. Afterwards, all groups were put together and each group presented

its results, on which the other groups commented.

34

To get women to express their points of view, the groups of women and men were
separated. The discussions were facilitated by ULPC, ACOD, OHVN, and ICRISAT

technicians, Deborah (a helper), and me.

Data Documentation and Analysis

The seed collection and the entry of identification information for each seed
sample and physical purity data were done by the driver (Sidy Dembele) and me. The
surveys were conducted by four enumerators (two men and two women), closely
supervised by me. I was an enumerator when needed4 and I checked all data quality in the
field and after data entry. The data collection as well as entry from the surveys, and lab
analyses was conducted by one person supervised by me.

The lab analyses data included the percentage of germination, percentage of
impurities (by weight), and the number of grain and seedling infected for each seed
sample. An example of the survey is presented in appendix 5. The quantitative data
collected in the survey included socio-economic characteristics and cropping system
traits of the households. Qualitative data collected in the survey included gender, zone,
seed sources, village, seed storage, and variety type.

Data were first registered on sheet of papers in the lab and on questionnaires for
the surveys in villages. Then all the data was entered in Microsoft Excel spreadsheets.
The reporting was done using the program Microsoft Word for reporting and the analysis
was done using the programs of Excel and Statistical Analysis Software (SAS 9.1.2).

Descriptive statistics were used, including frequency analysis, percentages, mean values

 

4 When we found that there many farmers in the sample ready to be interviewed quickly so that they could
leave for other household needs (farm work. trip, etc...) or other reasons, in order to get everybody
interviewed on time I served as an enumerator. This situation was very common.

35

and standard deviation. The analysis of variance and other statistical tests were done
using SAS Generalized Linear Model (GLM) using the CONTRAST statement for the
lab analyses and survey data. The GLM PROCEDURE with CONTRASTS enables one
to perform custom hypothesis tests (SAS Inc 1999).

The GLM could not be conducted for seed source data, as market sourced seed had a
very small sample size compared to own production seed; instead trends were discerned
through descriptive statistics. To assess significance, we used or: 20%, i.e., a 20%
significance level. This is an appropriate level of significant for on-farm research, given
the inherent high variability and importance of discerning trends to elucidate the

complexity of seed systems (Manderscheid, 1965).

36

RESULTS

Surveys

Since the survey sample was chosen based on the household’s willingness to
participate in the research activities, the results presented here did not include the
opinions of the farmers who did not want to participate. However, they are representative
of the farmers interested in research activities in the villages where the research was
conducted. The fact to not agree to participate in research activities, the nonparticipating
households might have different opinions about research those who were willing to

participate.

Socio-economic characteristics of the farmers

Consistent with the information obtained from seed collection, the survey found
that groundnut is the crop that both men and women are involved in growing, In Dioila,
more men (42%) seem involved in groundnut farming than in Mande (14%) (Table 1-4).
The middle age (30 to 50 years old) growers’ percentage is higher for groundnut, while
for sorghum and millet, the majority of growers interviewed are old (>50 years old). This
result can be explained by the fact that millet and sorghum are basic food crops and are
grown on the family field. Usually the head of the household represents the grower of
family crops.

At both sites, the majority of groundnut production is within fields managed
individually (80%) rather than the family collective fields (Table 1-4). About 90% of
sorghum and millet production is based in the family fields. In the majority of cases,
seeds used to plant personal fields are provided by the grower him or herself. Seeds for

the family field, on the other hand, are provided by the head of the household.

37

Groundnut field size and the number of workers per groundnut field are
substantially less (0.65 ha, and l to 3 people per groundnut field) compared to sorghum
and millet (6 to 7ha, and 3 to 4 people per field). A major portion of the farm, and family
labor, is devoted to family crop production.

The proportion of growers who received some education is also low among those

who grow groundnut, compared to sorghum and millet producers in both zones.

38

Table 1-4: General information (gender, age, information on the field) of the farmers who
participated in the research in 2007 in the two zones: Dioila and Mande

 

Groundnut Millet Sorghum
Dioila Dioila Dioila Mande
n=53 Mande n=79 n=25 Mande n=15 n=79 n=75
Gender of growers (% in the sample per gender)

 

Men 42 14 96 87 96 97
Women 58 86 4 13 4 3
Age group* of the growers (% in the sample per age group)

Young 34 37 16 27 15 13
Middle age 36 42 40 33 39 40
Old 30 21 44 40 46 47
Field type (% in the sample per field type)

Family

field 26 14 92 93 96 88
Individual

field 74 86 8 7 4 12
Responsible of seed provision for the field (%)

Head of

family 19 5 48 54 63 47
Person

responsible

for farm

work 34 32 36 13 27 36
Grower

him/herself 47 63 16 33 10 17
Average

field size 0.6 0.7 3.8 3.4 6.6 5.7
(SE) (0.1) (0.1) (0.8) (0.8) (0.7) (0.6)
Average #

of workers

in the field 3 1 7 5 7 4
(SE) (0.4) (0.1 ) (1.4) (1.2) (0.8) (0.6)
Education

(%) 32 28 40 73 49 45
Level of formal education reached by the growers who received some

schooling

Elementary 12 27 3O 27 23 47
Middle

school 29 5 10 9 16 15
Local

language 29 41 60 54 51 15
Other 30 27 0 10 10 23

 

*Head of household less than 30 years old: young, between 30 and 50 years old = middle
age, and above 50 years old = old.

39

Table 1-5: Information on the participating household cropping systems in Dioila and
Mande

 

 

Dioila n=80 Mande n=74
Cropping systems n % sample n % sample
Sorghum only 15 18.7 23 31.1
Millet only 5 6.3 2 2.7
Millet & sorghum 24 30 1 I 14.9
Groundnut &sorghum 10 12.5 16 21.6
Groundnut &millet 2 2.5 3 4
All three 24 30 17 23
Other 0 0 2 2.7

 

Sorghum as a sole crop (31%) is the dominant crop planted in the family fields in
the Mande zone, while the association of millet and sorghum was 30%, and similarly the
intercrop of three crops was 30% of family fields in the Dioila zone (table 1-5). This
result is consistent with the high proportion of millet seed samples collected in Dioila and

moderate levels of millet seed collected in Mande (see table 1-6 below).

Description of the seed samples used for the test

We examined seed quality parameters in relationship to location (zone and
village), variety type, gender of farmers, seed source and storage methods.

Table 1-6: Seed samples collected from farmers in Dioila and Mande on May 2007 for
lab-analyses of farmers’ seed quality parameters

 

 

Groundnut Millet Sorghum Total/zone
Zones n % n % n % n %
Dioila 74 33 44 20 104 47 222 100
Mande 89 44 17 8 97 48 203 100
Total 163 38 61 14 201 47 425 100

 

As shown in table 1-6, more groundnut samples were obtained in Mande than in
Dioila, indicating that groundnut was grown to a greater extent in Mande (53% of the

total groundnut samples and 44% of the samples from Mande) compared to Dioila (47%

40

and 33% respectively as in Mande) . For millet, the reverse was observed, where millet

was very important in Dioila (72% of millet samples and 20% of Dioila samples)

compared to Mande. This is consistent with the drier conditions in Dioila, an

environment that millet is adapted compared to sorghum. The two zones are almost the

same in terms of sorghum production. This confirms that both zones are major sorghum

producers.

Table 1-7: The number of seed samples collected in eight villages representing two zones

in Mali; Dioila and Mande.

 

 

Zones/Villages Groundnut Millet Sorghum Total
n % n % n % n %
Seribila 23 36 17 26 25 3 8 64 100
Mangnambougou 6 I6 12 32 21 52 38 100
Dioila Wakoro 20 3 1 4 6 40 63 64 100
Wobougou 25 46 l 1 20 1 8 34 54 100
Gonsolo 3 5 59 0 0 23 41 59 100
Kenioro 8 22 0 0 29 78 3 7 1 00
Mande Siranikoro 20 45 2 4 2 1 5 1 44 100
Siby 26 40 15 23 24 37 65 100
Total 163 3 8 61 14 201 47 425 100

 

Note: %= the percentage within village per seed samples, n= number of seed samples

collected

In Dioila, when we look at the percentage of seed samples collected in each of the

4 villages per crop, except for Wobougou, sorghum is the crop with highest percentage in

the samples in all the villages: 63% of the samples in Wakoro, 52% of the samples in

Mangnambougou and 38% in Séribila. Groundnut has the highest percentage (46%) in

Wobougou and is the second crop in Séribila and Wakoro. Millet shows up in second

place in Mangnambougou with 32% of the samples and in the third in the other three

villages.

41

In Mande zone, sorghum is the first crop in Keniero (78%) and Siranikoro (51%),
Groundnut is the first crop in Gonsolo (59%) and Siby (40%). In contrast to the Dioila
zone, we did not get any millet samples in Gonsolo and Keniero.

Table 1-8: Number of seed samples collected per gender and per crop (groundnut, pearl
millet, and sorghum) in Dioila and Mande for farmers’ seed quality assessment study

 

 

Gender Groundnut Pearl Millet Sorghum Total

n % n % n % n %
Men 57 35 61 100 196 97 314 74
Women 106 65 0 0 5 3 l l 1 26
Total 163 100 61 100 201 100 425 100

 

From table 1-8, it is clear that sorghum seeds (97% from men) and millet seed
(100% from men) are not crops that seeds are produced by women. Groundnut, on the
other hand, is the crop where both men and women are involved in the production, but
65% of the seed samples were from women which is consistent with it being a crop
produced mainly by women. This finding supports my choice of groundnut as crop to get
women involved in my research and confirms why women’s participation in millet and
sorghum plant breeding participatory research has been limited to variety evaluation
visits, and conducting culinary trials in southern Mali.

Table 1-9: Number of seed samples collected per type of varieties (modern versus local)
for all three crops in Dioila and Mande

 

Variety type Groundnut Pearl Millet Sorghum Total
Local 151 60 159 370
Modem 1 2 1 42 55

 

 

Looking at the variety types, local varieties representing 87% of all seed sample
collected are dominant for all three crops. Except for sorghum, where the modern
varieties represent about 21% of sorghum seed samples collected, there were apparently

few modern varieties of groundnuts or millet adopted in these zones. Millet had almost

42

no modern varieties. For testing the hypothesis about comparing modern and local
varieties, I therefore chose to focus on sorghum only.

Table l-lO: Number of seed samples collected per seed sources for all three crops
(groundnut, pearl millet and sorghum) in both zones: Dioila and Mande

 

 

Seed sources Groundnut Pearl Millet Sorghum Total
- Own production 158 58 195 I 411
Market 4 2 4 10
Exchange 1 1 l 3
Gift 0 0 l 1

 

For seed sources, own production represented about 97%, and market sources
were ~2% of the total. For hypotheses testing, I focus on two sources only--own
production and market--for the comparison because the sample size for the other sources
is too small. Also I want to know seed quality for market seed versus own production in
anticipation of making recommendations for promoting seed marketing.

Table 1-11: Percentage of seed samples sourced in market versus farmers own produced

seed per zone (Dioila and Mande) and per crop type (groundnut, pearl millet and
sorghum)

 

 

Groundnut Pearl Millet Sorghum total
Own Market Own Market Own Market Own Market
Pr Pr Pr Pr
Dioila 93 5 95 3 97 2 95 4
Mande 1 00 0 100 0 1 00 0 1 00 0

 

Own Pr = farmer produced seed

Self provision of seeds was relied on for more than 90% of seeds in all three
crops, thus own production was the most common source of seeds used in the study areas
(Table 1-11). Market sourced seeds for groundnut (5%) was slightly higher than for the

two other crops.

43

There were many storage methods. To make the analyses feasible, 1 grouped them
into the following simplified categories: seed stored in grains versus seed stored in
panicles or pods.

Table 1-12: Number of seed samples collected, by seed storage method for sorghum,
millet and groundnut per zone

 

 

Sorghum Millet Groundnut
Grains Panicles Grains Panicles Grains Pods
Dioila 54 50 41 3 2 72
Mandé 17 80 1 l 6 0 89
Total 71 l 30 52 9 2 l 61

 

From table 1-12, in Dioila zone, farmers stored most of their sorghum seeds as
grain (52%) rather than as panicles. In Mande, storage in panicle was the most common
method: 83% of sorghum seed samples were stored in panicles compared to 17% as
grain. For millet, the storage of seed as grain was the common practice in both zones.
Storage in pods (unshelled) was the most common storage method for groundnut seeds

across zones.

Lab analyses

The following parameters were assessed to evaluate seed quality: physical purity
(percentage of the weight of impurities), physiological quality (germination percentage)
and health status (% anthracnose for sorghum alone before germination test and the
percentage of seedling infection with germination test) (Table 1-13).

Hypothesis 1: Market-sourced seeds were hypothesized be higher quality than farmer’s

own seed production.

The general descriptive statistics (mean, and standard deviation) presented allow a

comparison between the two seed sources, market and own production. For sorghum and

44

millet, the mean level of impurities observed for all sources was below 10%. For
groundnut, the mean level of impurities was slightly above 20% for own production and
slightly less than 20% for market-sourced seed. In both cases, own production and
market, the level of impurities was above the level allowed for certified seeds (4%)
(Ntare et al., 2008).
All seed samples were found to have an 80% or greater germination rate, which is

sufficient to be high quality seed (see table 1.1).

In terms of seed infection, a key difference was observed between own production
and market source in groundnut (figure 1.3). Market-sourced seeds had less infected

seeds (18%) compared to own production (22%).

 

 

 

 

4o , 7
El own production
35 I Market 7
30
6 25
E
a) 20
U)
59
r: 15 ,
a)
8
o 10
o.
O I 7'71 l
Sorghum Millet Groundnut
Croptypes

 

Figure 1—3: Percentage of seed impurities weight by crop for market versus farmers own
produced seeds of seed samples collected in Dioila and Mande. The bars represent the
average weight (means) and lines above the bars represent the standard deviation from
the mean

45

 

120.00

100.00

60.00

40.00

Percentage germination

20.00

0.00

 

80.00 '4

 

 

Sorghum

[:1 Own production}
I Market

 

III

Millet Groundnut
Crops

 

Figure 1-4: Percentage of germination by crop for market-sourced versus farmers’ own-
produced seeds of seed samples collected from farmers in Dioila and Mande. The bars
represent the average weight (means) and lines above the bars represent the standard

deviation from the mean

40. 00

35.00

30.00

20.00

15.00

10.00

Number of infested plants in %

5.00

0.00

 

Figure 1-5: The percentage of infected seedlings for each of three crops for market-
sourced versus farmers’ own-produced seed of seed samples collected from farmers in
Dioila and Mande. The bars represent the average weight (means) and lines above the

25.00 -

 

fi , s

Sorghum

iEI Own production
II Market

Millet Groundnut

crops

bars represent the standard deviation from the mean

46

 

 

The infection of sorghum seed was the opposite of that observed for groundnut, i.e.,
the market-sourced seeds had the higher infection rate (figure 1-4). Overall, the infection

level for sorghum was low, less than 10% for both sources.

Hypothesis #2: The table 1-13 below presents data comparing seed quality of
modern varieties in sorghum to local varieties. No effect of variety was found on
sorghum seed quality.

Table 1-13: Seed quality characteristics are presented for modern and local sorghum

varieties, along with standard Error of the mean (SE) and a GLM contrast to test if
modern versus local varieties are significantly different.

 

 

Statistic for variety Seed quality parameters
types
WOI Anthracnose Germination Infection
Infection
(%)

Local
n=159 Means 2.7 2.8 88 7.1

SE 0.2 0.2 7.0 0.6
Modern
n=42 Means 2.5 3.4 88.0 8.1

SE 0.3 0.5 13.5 1.2
Contrast P-value 0.78 0.52 0.91 0.80
modern vs
local

 

WOI refers to the percentage of the weight of impurities found in a seed sample
SE refers to standard error of the mean

These results are not conclusive due to high variability they do not provide
evidence that seed quality of modern varieties is different from local varieties.
Hypothesis #3: Seed quality of sorghum samples was evaluated to determine if the
seed location was an important determinant. It was found that seed from Dioila had
slightly higher rates of impurities, anthracnose infection and seedling infection compared

to Mande(Table1-14).

47

Table 1-14: Seed quality characteristics for Dioila and Mande sorghum varieties, along
with standard error of the mean (SE) and a GLM contrast to test if quality characteristics

of seed from Dioila are significantly different from those from Mande.

 

Statistic per zone

Seed quality parameters

 

WOI Anthracnose Germination Seedling
Infection
(%)
Dioia n=104 Means 3.7 88.1 9.6
SE 0.4 8.6 0.9
Manden=97 Means 1.5 88.2 5.9
SE 0.1 9.0 0.6
Contrast P- 0.001 1 0.95 0.0002
modern vs value
local

The seed quality response was further examined by village, where it was shown

that the highest level of impurities was found in Wakoro, Dioila zone. The highest

percentage of anthracnose infection (9%) and the highest percentage of general infection

(16%) were found in Mangnambougou, another village in Dioila zone (table 1-15). Thus,

these results are consistent with sorghum seed quality being generally better in Mande

compared to Dioila.

Table 1-15: Seed quality analyses results for sorghum are presented by village of
production Mean, and Standard Error presented in parentheses, to identify which villages

 

 

have better quality seed
% impurity % Anthracnose infection % infected plants

Mangnambougou 3.2 (0.7) 9.0 (1.9) 16.0 (3.5)
Séribila 2.6 (0.5) 2.5 (0.5) 7.5 (1.5)
Wakoro 5.0 (0.8) 1.1 (0.2) 6.3 (1)

Wobougou 2.6 (0.6) 4.6 (1.1) 12.1 (2.9)
Gonsolo 1.2 (0.2) 1.1 (0.2) 6.6 (1 .4)
Keniero 1.4 (0.2) 1.3 (0.2) 4.5 (0.8)
Siby 1.1 (0.2) 4.1 (0.9) 6.7 (1.5)
Siranikoro 2.7 (0.5) 1.3(0.3) 4.7 (0.9)

 

To provide insight into these results, I looked at the two extreme values per seed

parameter that tested significant. Table 1-16 shows this information.

Table l-16: Factors contributing to high values in impurities observed among sorghum
seed samples, including varieties and villages where seed were produced.

 

 

Extreme Sample Variety Name Village
value number
42 l 5 Soguerekou Wakoro
Dioila 40 1 4 Algerien Wakoro
%WOI Mande 1 0 1 72 Tiemarifing Keniero
10 126 Kendebilen Gonsolo
% 42 61 Bandoka Wobougou
Anthracnose Dioila 24 79 Algerien Magnambougou
Infection Mande 44 1 30 Tieblen* Siby
1 8 1 10 Dorongonnikalan Siby
% general 40 79 Algerien Magnambougou
infection Dioila 35 61 Bandoka Wobougou
Mande 27 1 16 SUMALEMBA* Gonsolo
20 1 13 Dorogonnikalan Siby

 

The varieties with * are modern varieties.

Table 1-17: Seed quality characteristics for Dioila and Mande for pearl millet and
groundnut, along with standard Error of the mean (SE) and a GLM contrast to test if seed
fi'om Dioila and Mande are significantly different.

 

Statistic per zone

Seed quality parameters

 

WOI Germination Infection
Dioila %
n= 44 Means 4.0 81.0 3.0
Pearl SE 0.6 12.3 0.4
Millet Mande n=17
Means 1.3 86.0 2.0
SE 0.3 21.0 0.5
Contrast Dioila P-value 0.0379* 0.0506* 0.0418*
vs Mande
Dioila %
Groundnut n= 74 Means 17.0 85.0 16.0
SE 2.0 9.9 1.9
Mande n=89
Means 26.0 76.0 24.0
SE 2.8 8.1 2.5
Contrast Dioila P-value <.0001 <.0001 0.0046

vs Mande

 

For millet, there was evidence that seed quality varied by zone, with higher

quality seeds being produced in Mande (tablel-17).

49

In contrast to the observations on seed quality for cereals, the opposite trend was
observed in groundnut. Dioila was associated with higher seed quality. Seeds from Diola
had over 17% weight of impurities, 85% germination rate and 16% infection level, which
were better levels than groundnut seed from Mande with respectively 26%, 76% and 23%
(table 1-17).

Hypothesis # 4: Comparing storage methods

For sorghum, comparisons were made in relationship to three important methods
of storage: grains, sheaves, and panicles not attached together.

Table l-18: Seed quality characteristics per storage methods commonly used by farmers

for each crop, along with Standard Error (SE) and a GLM contrast to test if quality
measures for seeds stored in panicles, sheaves, and grain are significantly different.

 

Seed quality parameters

 

WOI Ant Infect Germ Infection

Grain n=72 %

Means 3.9 2.3 90 8.4

SE 0.4 0.3 10.6 1.0
Sheaves n= 64

Means 2.5 5.2 87 10

SE 0.3 0.6 10.9 1.2
Panicles n=65 %

Mean 1.4 1.4 88 5.1

SE 0.2 0.2 10.9 0.6
Contrast grain vs. P-value 0.0089 0.19 0.180 0.38
sheave and panicle
Contrast sheave vs. P-value 0.191 0.0001 0.72 0.0001
Panicles

 

Ant Infect refers to Anthracnose Infection, Germ refers to Germination rate

Sheave is a sorghum storage method where farmers select the good panicles during or before harvest and
attach them together to form a sheave that they hang inside the granary or the kitchen and/or hang it on a
tree in the courtyard.

The results for sorghum show that the difference between seeds stored in grains
and those stored in panicles (sheaves and single panicles) is highly significant for WOI

only. Seed stored as grain, with about 4% W01, has significantly more impurities than

50

seed stored as panicles (2%). For anthracnose, the trend was reversed: seeds stored as
grain had 2.0% infected grains, whereas seed stored in the panicles had 3.3% infected
grains. The same trend appears with germination rate: seed stored as grain has a 90%
germination rate, compared to seed stored in panicles with 87.5%. The comparison
between sheaves and single panicles also showed the all quality parameters were
influenced by storage method, with the exception of germination rate (Table 1-18). Seeds
stored in single panicles had higher quality parameters (1.4% WOI, 1.4% Ant-infection,
and 5.1% total infection) compared to seeds stored in sheaves (with 2.5%, 5.2 and 10%,
respectively).

Groundnut seeds stored in pods, in a bag, had slightly lower impurities, with 20%
W01, compared to 25% in pods stored in a granary (table 1-19). Seeds stored in pods in
a bag had an 83% germination rate and 18% infection rate, compared with a 73%
germination rate and 25% general infection for seeds stored in pods in granary

For millet, there was no significant difference between seeds stored as grain and
seed stored as panicle, for all quality parameters assessed (table 1-19). However, the

sample size for panicle storage was low.

51

Table 1-19: Seed quality characteristics per storage methods mainly used by farmers for
groundnut and pearl millet, along with standard error of the mean (S E) and a GLM
contrast to test if seeds stored by different method are significantly different.

 

Seed quality parameters

 

WOI Germination Infection
Pods in bags n=97 %
Means 20.4 83.0 17.7
SE 2.1 8.5 1.8
Pods in granary n=64
Groundnut Means 24.0 75.0 25.0
SE 3.0 9.4 3.1
Contrast pods in bags P-value 0.0925 0.0002 0.0068
vs. pods in granary
Panicles n=9
Means 2.0 84.0 2.4
Pearl SE 0.7 28.0 0.8
Millet Grain n=52
Means 3.8 82 2.5
SE 0.5 11.4 0.3
Contrast Panicles vs P-value 0.354 0.5324 0.898

grain

Hypothesis # 5: Comparing groundnut versus millet and sorghum seeds

Table 1-20: Seed quality characteristics are for sorghum, pearl millet and groundnut seed
samples along with standard error of the mean (SE) and a GLM contrast to test if seed of
the three crops are significantly different from each other.

 

Seed quality parameters

 

WOl Germ Infection
%
Sorghum Means 2.7 88.0 7.8
n= 201 SE 0.2 6.2 0.5
Millet n=6l Means 3.6 82.4 2.5
SE 0.5 10.7 0.3
Groundnut n=163 Mean 21.8 80.4 20.4
SE 1.7 6.3 1.6
Contrast groundnut vs Millet & P-value <.0001 0.0005 <.0001
Sorghum
Contrast Millet vs and Sorghum P-value 0.5 0.0026 0.001 1

 

52

To evaluate biology of crop effect on seed quality, the effect of legume versus
cereal crop types was tested by comparing seed quality in groundnut, a legume crop, to
that in sorghum and millet, cereal crops. A second comparison was done between the
two cereal crops. The results of the first comparison showed that there were highly
significant differences between legume and cereal species (P-values for the three
parameters were respectively <0.0001, 0.0005, <0.0001). Groundnut had 22% W01, an
80% germination rate and a 20% seed infection rate. This is indicative of considerably
more seed quality concerns for groundnut than sorghum and millet with 3%, 85%, and
5% respectively of WOI, percentage germination and infection.

The second comparison between sorghum and millet showed that germination
level for sorghum seeds was better at 88% compared to millet at 82% (P-values for the
germination rate was 0.0026 ). However, millet had a 2.5% infection level and thus much
less seedling infection potential compared to sorghum seeds at 7.8% infection (P-value
for infection was 0.001 1).

Hypothesis # 6: The influence of gender on seed quality could only be assessed for
groundnut, which was grown by both men and women.
Table 1-21: Seed quality characteristics per gender for groundnut seeds collected from

farmers in Dioila and Mande zones along with standard error of the mean (SE) and a
GLM contrast to test if seed from men and women are significantly different.

 

Seed quality parameters

 

WOI Germination Infection
%
Women n=106 Means 24.0 78.0 23.0
SE 2.3 7.6 2.2
Men n= 67 Means 17.0 85.0 16.0
SE 2.1 10.4 1.9
Contrast Men vs Women P-value 0.0006 0.0007 0.0101

 

53

The results of these comparisons showed that there are significant differences
between men and women in term of groundnut seeds management. For all three quality
parameters tested: WOI, germination rate and infection, men‘s seed samples showed to

be better quality than women’s seed samples.

Farmers seed storage from survey

The survey results show that storage in pods and panicles is the most common
seed storage practice for all three crops; and this was consistent across zones (table 1-22).
Comparing the two zones, the amounts of seeds stored in pods/panicles in Mande are
higher for all three crops (82%, 67% and 81% respectively for groundnut, millet and
sorghum).
Table 1-22: Seeds storage method used by farmers in Dioila and Mande, for groundnut,

millet, and sorghum observed during farmers seed quality assessment study done from
May to December 2007 in southern Mali

 

 

Groundnut Pearl Millet Sorghum
Diola Mande Diola Mande Diola Mande
n=53 n=79 n=25 n=15 n=79 n=75
Storage method (% in the sample of the survey)
In grains 6 0 44 33 24 1 l
Pods/panicles 77 82 48 67 75 8 1
No Response* 21 18 8 0 l 8
Pods/panicles storage method (% of people who stored in pods/panicles)
ln bulk in granary 34 34 24 13 25 23
In bags stored in granary 13 16 12 20 7 14
Sheaves stored outside 0 0 20 20 22 34
Sheaves stored in
granary 0 0 20 40 22 29
No Response* 53 50 24 7 24 0

 

* The survey participation consent statement allowed the respondent to not answer questions if she/he
doesn’t have answer for or doesn’t fell comfortable in answering.

Bulk pods or pods in bags within a granary are common practices for groundnut
storage in both zones (table 1.22). Millet panicles are stored equally outside and in

granaries in Dioila, while in Mande sheaves are in the main stored in granaries (40%).

54

Farmer seed storage issues

As presented in the table 1-23 below, during storage, farmers are concerned about
different issues depending on the crop type. For groundnut, farmers worry mostly about
insect attack on seeds. In Dioila, 57% of groundnut growers were concemed about insect
attack, whereas 68% of farmers in Mande were concerned. Insect attack is often followed
by further degradation, a concern of 32% of farmers in Dioila, and a related concern
about humidity was expressed by 10% of farmers in Mande. During millet storage,
farmers worry mostly about temperature because very high temperature can kill the
embryo, and more so in Dioila (76%) than in Mande (40%). For sorghum, insect attack is
a critical factor of concern during seed storage; 76 to 80% expressed this concern at both
sites. Other physical damage following insect attack was also of concern in sorghum seed
storage, along with being an issue in groundnut.

Table1-23: Factors of concern for farmers during seed storage of the groundnut, millet,

and sorghum in Dioila and Mande identified during the farmers seed quality assessment
study done from May to December 2007 in southern Mali

 

 

Groundnut Pearl Millet Sorghum
Diola Mande Diola Mande Diola Mande
n=53 n=79 n=25 n=15 n=79 n=75
Factors of concern during seed storage
Temperature 2 6 76 40 4 1 5
Humidity 2 10 0 8 6 17
Insect attack 57 68 0 0 76 80
Other physical degradation 32 8 40 33 13 29
Prevention of seed loss during storage
keep seed in pod/panicles 28 l 4 0 6 4
Chemical treatments 0 63 44 l 3 47 1 5
Regular checking during
storage 13 2 8 20 1 l 21
Other 21 5 20 27 16 12

 

The percentages (%) in this table indicate the frequency with which a quality parameter was cited by
farmers.

55

To address the concerns described, farmers use different methods. For groundnut,
the most popular practice was to store seed in pods (unshelled) in Dioila (28%); while in
Mande seed treatment was the key practice for seed protection by groundnut famers
(63%). In Dioila, both millet and sorghum are protected using seed treatments, reported
by 44 to 47% of farmers respectively. In Mande, other prevention practices were
reported by 27% of farmers for millet, and regular checking during storage was reported

for sorghum (21 %).

3-3-2 Seed treatment during storage at farmers’ level

Table 1-24: Seed treatment used by farmers in Dioila and Mande during seed storage for
the groundnut, millet, and sorghum documented during farmers seed quality assessment
study done in 2007 in southern Mali

 

 

Groundnut Pearl Millet Sorghum

Diola Mande Dioila Mande Diola Mande

n=53 =79 n=25 n=1 5 n=79 n=75
Treatments during
storage (Yes) 28 77 44 27 34 19
No 51 6 44 40 42 45
No response 21 38 12 33 24 36
Type of treatment (% of people who treated seeds during storage)

n=15 n=6l n=1 1 =4 n=27 n=14
Spraying insecticide
liquid 7 3 36 25 7 0
Use of insecticide in
small pieces 53 87 36 25 67 36
Use of insecticide in
powder 27 2 18 25 18 28
Traditional treatment 7 2 0 0 0 0
Others 6 6 10 25 8 36
Time of the treatment
Beginning of storage 67 84 64 25 81 28
In middle of the storage 0 0 0 0 4 36
Others 33 16 36 75 15 36
#of treatment
One treatment 73 74 45 50 74 78
More than one 7 8 19 25 4 14
No Response 20 18 36 25 22 8

 

56

Seed treatment during storage is not common practice for millet and sorghum in
both zones. For groundnut, 77% of the farmers interviewed in Mande said that they
treated their seeds during storage, while only 28% did in Dioila zone. Of the people who
did treat their seeds, most used insecticide in small pieces for all three crops. Here again,
87% of Mande farmers have used this product for groundnut and 67% in Dioila did for
sorghum. Except for millet in Mande, most of the treatments were done at the beginning

of the storage.

Farmers’ assessment of seed quality

Tablel-25: Farmers’ seed quality evaluation as assessed through survey questions
conducted before and after planting on December 2007 in the eight villages representing
Dioila and Mande zones in Mali

 

 

Groundnut Millet Sorghum

Dioila Mande Dioila Mande Dioila Mande
QualityI parameters n=53 n=79 n= 25 n=15 n=79 n=75
Before planting
Appearance 72 90 52 60 63 77
Past experiences 1 7 27 4 0 21 28
lnforrnation from other '
growers 28 25 20 20 33 28
Seed coat integrity 53 78 56 67 40 76
% impurity 40 46 32 40 34 37
Seed shape 57 54 28 60 37 53
Seed color 41 37 24 33 30 29
Seed size 55 35 32 47 27 47
Insect attack 62 44 56 53 48 56
Presence of other
grains 4 16 0 7 4 1
After planting
Germination rate 77 66 72 80 71 83
Vigor of seedling 9 24 4 40 1 l 24
#of healthy plants 6 35 4 20 5 39
Others 0 4 4 0 4 4

The percentages (%) in this table indicate the frequency with which a quality factor was cited. where
farmers could report more than one seed quality attribute

57

The results in this table show that farmers put different weight on quality
parameters depending on the crop type. Before planting, appearance is the most important
quality parameter for groundnut and sorghum in both zones (table 1-25). In Mande,
almost all farmers (90%) surveyed prioritized the appearance of groundnut seed before
planting. For both sorghum and groundnut, appearance was followed by insect attack
(48% sorghum and 62% groundnut) in Dioila, and seed coat integrity (76% for sorghum
and 78% for groundnut) in Mande. For millet, the most important quality parameters
before planting were seed coat integrity and insect attack (56%) in Dioila and seed coat
integrity (67% ) followed by appearance and seed shape (60%) in Mande (table 1-25).
This changed after planting. The percentage of germination was the most popular seed
quality factor later in the season and was reported by ~ 75% of farmers for all three crops,

across ZOIICS .

58

DISCUSSION

In Mali, it is a general assumption by researchers and policymakers that farmer-
produced seed is of inferior quality compared to formal sector and research produced
seed (Diakité, 2004, Yapi et al., 2000). According to Diakité (2004), the wide use of local
sector seeds constitutes one of the major challenges facing sorghum production in Mali.

My research tested this assumption for different seed sources, storage types and crops.

Cropping system characterization

As expected, cropping system patterns followed agroecological zones: millet was
grown primarily in Dioila, which has a drier climate than the Mande region, reflecting the
adaptation of millet to low rainfall conditions. A majority of the groundnut samples were
also found in Mande, about 55% more of the sample compared to Dioila (table 1-6). This
finding is somewhat surprising, as groundnut generally requires more rainfall than is
typical of Dioila. The higher prevalence of groundnut in Mande may also reflect the role
of alternative cash crops in Dioila, as cotton has been grown for decades in this zone
supported by the Malian cotton company (CMDT) , and this may reduce farmers’
incentives to grow groundnut as a cash crop (Boughton, 1994; ICRISAT, 2002).

There are also socio-economic factors that influence cropping patterns. According
to Malian tradition, the male head of the household is responsible to ensure the staple
grain food of sorghum and millet (Becker, 2000), whereas women are expected to take
care of the ingredients of the sauce, including groundnut. As a result, men usually grow
millet and sorghum, while married women grow sauce crops such as groundnut and okra

(Brock 2007). The data from our survey on field size, the age group of the grower, the

59

field type and the number workers per field all were consistent with the importance of
sorghum and millet as family staple crops (table 1-4).

The literacy rate for producers of groundnut was lower than the rate for farmers
that primarily produced millet and sorghum; this is presumably due to the vast majority
of cereal producers being men, who have a higher literacy than women who made up
about half of the groundnut producers. Based on these results, we are not surprised that
on-farm research in this area, which has focused primarily on cereals, has had low

participation of women (Broek, 2007).

Farmers’ seed sourcing

The informal seed sector, particularly the ‘own-produced seed,’ is the most
important seed source for all three crops. As shown in table 1-11, own-produced seeds
represented more than 90% of seed samples collected. This result reflects the trend of
seed provision in the research areas. Siart (2008), in her research on strengthening local
seed systems conducted in the same zone, found similar seed source results. Diakité
(2003), in his study of promoting improved varieties diffusion through seed sale, showed
that seeds, particularly sorghum seeds, in the research areas have more social values than
market value. Seeds are there to consolidate the deep social relationships within a
community, not to be sold in much of Mali (Siart, 2008, Broek, 2007, Diakite 2003). And
he concluded that any sustainable and efficient action toward improving farmers’ seed
provision in these two zones should take into account the realities of the informal seed
sector.

Farmers reported in Mande that no seed was sourced from the market (Table 1-

11). This was surprising, as NGOs and researchers have been promoting seed fairs in this

60

region, and this should have augmented availability of seeds in the samples from a range
of sources. For example, Siart (2008) documented seed fairs held in 2005 in Mande
where more than 90% of the participants were happy with the seed quality and about 80%
found that the prices were affordable. We have not found an explanation for this absence
of market-sourced seed in our sample, although it should be taken into consideration that
purchasing seed of staple crops is not culturally seen as appropriate behavior (Siart, 2008;
Brock, 2007; Diakite, 2004). Farmers may be unable to produce sufficient seeds, but for
socio-cultural reasons, they do not want to be seen buying them.

Groundnut seed was sourced to a limited extent (5%) from markets, higher than
the other crops. This result is not surprising because the biology of groundnut is
characterized by high germination and infection rates, as well as a high oil content, which
leads to spoilage and makes seed storage difficult (N ’tare et al., 2008). In addition to
biological reasons to purchase seed, farmers sometimes grow groundnut as a cash crop
and thus may be interested in expanded access. Further, groundnut seed if often lost
through consumption (e.g., during a social event such as a wedding) or due to emergency
cash needs (N ’tare et al., 2008). These findings in groundnut are similar to research on
bean seed systems in Eastern Afiica that indicated farmers often face multiple losses of
legume seeds, and intermittently rely on seed purchases to renew bean varieties (David

and Sperling, 1999).

Seed storage

A consequence of widespread reliance on own-seed production is that farmer

storage conditions are important determinants of seed quality at planting. Farmers have

61

developed specific storage techniques and practices that vary with culture and crop type
(Christinck et al., 2006).

Seed stored in panicles (or pods for groundnut), or as grain, were the two most
important storage methods used. Storage as grain was mostly used for millet. Farmers
report that they store sorghum seeds in pods/panicles to prevent damage caused by
insects and mice (N ’tarc et al, 2008, Siart, 2008, Diakite, 2003). Similarly, all groundnut
samples were stored in pods to protect against pest damage (author’s survey data 2007).
Sorghum storage in the Mandc zone involved substantial storage in panicles, much more
so than in Dioila (table 1-12). This result agrees with a Siart (2008) study conducted on
sorghum in the two zones. It also provides evidence that improved, local practice and
traditional knowledge is widely understood and used in Mande but much less so in

Dioila, where there are many extension services (ICRISAT, 2002).

Seed Quality

Farmers want seeds of good quality and with the characteristics they need for
their particular agro-ecological conditions and objectives (ICRISAT, 2000; Diakite, 2004,
Chakanda, 2000). However, the key aspects of seed quality such as capability to
germinate, age, health status, and the effect of storage can be difficult to judge when
acquiring seeds. In other words, seeds are not transparent in term of quality (Badstrue,
2007). The only way to evaluate seed quality parameters is through testing. The three
seed quality parameters measured in our study - physical purity, germination rate, and
health status — were found to be high for all farmer produced seed. This held true for the
three crops studied, all of which met the seed quality standards set by the SSN and the

Harmonized Standard of “Comite Inter-etat de Lutte contre la Secheresse au Sahel”

62

(CILSS) countries for groundnut (N’tare et al., 2008, Diakite, 2003, Almekinders and
Louwaars, 1999). The only exception of was infection level, which was higher than the
recommended 0%, but we note that infection was less than 5%, which is on a practical

level what is achievable in seed production in Mali.

Seed quality for purchased versus farmers ’ own-produced seeds

Seed physical purity for sorghum and groundnut from purchased source seeds was
higher than that observed for farmer-produced seeds. This result was not surprising
because sorghum seed from a purchased source in the Dioila zone was from the
organization of cereals producers - OPCs (author’s survey data 2007), and the members
of these farmers organization receive training on seed production from ICRISAT and [ER
(Siart, 2008). In addition, during the meetings with farmers, they attested that appearance
of market seeds is better than own-produced seeds because sellers pay attention to
producing clean seed that to attract buyers. One farmer (from Keniero Mandé) said: “the
sellers do all they can to attract buyers whose only quality appreciation parameter is the
appearance at that very moment. They clean the seeds well and put them in clean bags or
boxes” (meeting with farmers, September 18th 2008).

In terms of germination rate and seed infection level, this study documented that
market-sourced seed was slightly better quality than fanner-produced seed for millet and
groundnuts (figures 1-3 and 1-4). This result for groundnut is not a manual by N’tare ct
al., 2008 concerning groundnut seeds from different sources. For millet, the results have
to be taken with caution because the sample size was too small.

For sorghum, in all cases the farmer-produced seed quality was high, with only

moderate or no reduction in quality compared to market-produced. The moderately

63

higher levels of infection in the market-sourced sorghum compared to farmer-produced
seed may be indicative of a major problem. Through the survey comments I found that
farmers report that market seeds of sorghum are varietally pure, but are often infected,
which leads to diseased seedlings (author’s follow-up workshops 2008). A farmer from
Gonsolo (Maude) said: “If we are running out of seeds, and cannot get it with a friend or
relative, that is when we use the market source, but seedsfrom market are infected and
we cannot tell at the time we buy them. When the infection starts, it is too late to replant
and we don 't know what to do. ” Another farmer from Siby said “Even if they look good
(clean), we don 't like market seeds because we don 't know the variety " (meeting of
September, the 17‘“ 2008).

This result should take seriously because it shows one of the obstacles to the
promotion of seed markets and the diffusion of modern varieties in Mali (Yapi et al.,
2000; Diakite, 2004). However, this finding is consistent with ICRISAT (2000)
statement that “ armers have selected seed from their cropsfor the next season. Their
knowledge has been accumulated over generations, and nobody is in better position than
they are to choose seed for them”.

To summarize, the results suggested that farmers are producing high quality seed,
which may reduce the incentive for farmers to purchase seed. There may be other
considerations beyond quality that influence the very low percentage documented here of
farmers sourcing seed from market. There are other factors than seed quality that could
explain reluctance to purchase seed.

0 The cultural view of seeds as a non-market good, as mentioned by Diakite,

2004, in his study on the use of sale as way to promote the diffusion of

64

improved varieties in Dioila, and Brock, 2007, field work paper on
sorghum production by women in both zones.

0 The lack of information on market seeds, as suggested by Sperling et al.,
2006 in assessing seed security in northern Mali, and Badstrue, 2007, in

her paper “The role of trust in the acquisition of seeds”.

Seed quality for modern varieties versus local variety seeds

The analyses of the three quality parameters for sorghum modern and local
varieties showed that there was no significant difference between modern and local
varieties with respect to these quality parameters. The study by Siart (2008) on
strengthening local seed systems in Dioila and Mande zones showed that under different
conditions (soils, rainfall, and temperature), farmers ‘local seeds achieve high
germination rates (>80%) and high physical purity (<2%)’. Yapi et al. (2000) also
showed that one of the reasons for low adoptions of modern sorghum varieties was
because they were not superior (adaptation, yield and taste) to local existing varieties.
Our findings suggest that the low yield potential of local sorghum varieties” which
Diakite, 2003, attributed to the wide use of farmers’ seeds, might be due to reasons other

than seed quality.

Seed quality by zone (Dioila versus Mande)

Farmers in Dioila are provided technical information from such organizations as
CMDT and many development NGOs such as ULPC, Netherlands Development
Organization (SNV) projects and others. Thus they have significantly more access to
information and training than in Mande (Siart, 2008), and we expected that Dioila

farmers would have access to and produce high quality seed. Surprisingly, sorghum seed

65

samples from Mande showed better quality characteristics compared to samples from
Dioila (table l-14). This indicates that traditional seed production and storage practices
(storage in sheaves) — which are practiced widely in Mande— are conducive to producing
high quality seed, at least in sorghum (table 1-12).

Another important factor may be the prominent presence of modern variety
S UMBA in Dioila compared to Mande. The improved variety called SUMBA was a major
source of infection in Dioila, more particularly in Mangnambougou collected seeds (table
1-15). According to farmers, the high infection of “SUMBA” is related to its early
maturity, which leads to it being harvested when it still raining and being stored at high
moisture content (unpublished data from follow-up workshops in Dioila, September
2008)

Overall, groundnut showed that reasonably high quality seed could be produced
by farmers, although physical purity and infection was higher (>10% ) compared to the
CILSS harmonized groundnut seed quality standards, which are respectively 4% and 2%
(N’tare et al., 2008). Overall, groundnut seed samples from Dioila were of better quality
than those from Mande (table 1-17). The germination rate was high from both zones, in
almost all cases greater than the CILSS standard of 70% for groundnut. This is surprising
in view of the high oil content and vulnerability to pests ofien found with this crop

(Tripp, 2000; N’tare et al., 2008, Nigam et al., 2004).

Seed quality by storage methods

Consistent with earlier studies conducted by ICRISAT in the research areas, we
found three major sorghum seed storage methods: grains, sheaves and single panicles

(Diakite 2003, Siart, 2008). Across a wide range of storage methods, the level of general

66

infection was above the SSN standard (>50/o). However, our research indicated that seed
stored as panicles and sheaves had better seed quality than seed stored as grain. Siart
(2008) found similar results, that seeds stored in the sheaf had better quality than seeds
stored as grain. However, there are tradeoffs in terms of storage methods. When we
compared the two storage methods for panicles (sheave and single panicles), we found
that single panicle seeds have fewer impurities (<2%), less anthracnose infection (<2%),
and less general infection (<6%) than sheaves. These results are opposite to the common
farmer assessment that sheaves were the best storage method for sorghum seed (author’s
informal discussion with farmers 2007). However, they were not surprising to farmers, as
they commented, when we returned to discuss initial laboratory analysis findings, that the
panicles of the sheaves are attached together just afier harvest (not dry); thus, the chance
of infection and contamination is relatively high compared to single panicles, which are
dried out before storage (author’s follow-up workshop in Dioila 2008) .

For groundnut, pods stored in bags were associated with. higher quality than pods
stored in the granary (Table l-l9). Again, the comments of farmers were instructive,
from our follow up visit to share the initial results. Farmers indicated that they check
regularly groundnuts in bags, removing damaged pods, and that bags are stored in a
chemically-treated place. The pods stored in granary, by contrast, are closed hermetically
from the beginning to the end of the storage period. The farmers open the granary only

when it is time to prepare seed for planting at the beginning of the growing season.

Seed quality in relationship with crop biology

A review by Tripp (2000) indicated that the ability of a crop to be maintained as

seed depends on seed characteristics; the biology of the species. N’tare et al., (2008)

67

pointed out that groundnut seed storage and handling are issues are of critical importance
due to the biological characteristics of the crop. In agreement with these reviews, my data
showed that groundnut samples had the poorest quality traits among the three crops
studied. Groundnut seed is characterized by high oil content, and is large sized, with an
easy detachable seed coat enhancing vulnerability to many post-harvest pests, and to

storage damage (N’tare et al., 2008).

Gender and seed quality in groundnut

Overall, higher quality was associated with groundnut seed collected from men
compared to women. This was not expected, considering the literature which indicates
women tend to be more knowledgeable about a crop that they are primarily responsible
for, such as groundnut (Saito et al., 1990). However during the follow-up workshop the
women told me that they did not give me their true groundnut seed because they did not
know that my research was important to them. Thus our research hypothesis could be true
if the women had provided their real seed during the seed collection.

This behavior of women can be a consequence of their weak participation in
research activities because they had no idea of what I was going to do with the seed
compared to men. At the workshop, they felt kind of sorry of not having given the good

seeds.

68

CONCLUSION

The results showed that farmers have similar concerns about seed quality as
researchers do, but the ways concerns were addressed were quite different. From the lab
analyses, we found that seed quality of farmer-produced seed was high, meeting CILSS
standards in almost all cases. The level of impurities was less than 10% for all sorghum
and millet seed sampled. The germination rates of sorghum and millet were over 80% in
all cases. Even for groundnut, which is known to be vulnerable to decay and poor quality
seed problems, the germination rate was consistently above the CILSS standard of 70%.

There were some concerns about physical purity and health status of groundnut
seeds, as some of the farmer-produced seed that was below the recommended
certification standards. These problems were serious in groundnut, particularly from the
Mande zone in farmers’ own-produced seeds. Since women are the most involved in
groundnut production, education of women on seed quality could help address this
problem. That will prevent them from doing the same thing of giving the bad seed as they
did in this research. Further, involvement of more women in participatory breeding and
research activities could provide long-term benefits for crop improvement in Mali.

The result of the comparison between modern local varieties showed that modern
sorghum varieties have lower seed quality compared to local varieties. Malian plant
breeders have developed new sorghum varieties that have high yield potential and are
early, but the negative traits related to seed quality need to be fully considered. Under a
system where farmers are saving seed, the fact that a variety is highly susceptible to

infection needs to be fully considered.

69

CHAPTER 2

IMPACT OF FARMERS’ SEED RECYCLING ON
UNIFORMITY OF SORGHUM VARIETIES IN SOUTHERN
MALI

INTRODUCTION

Maintaining varietal purity is associated with maintaining varietal identity and
thus the performance level of the variety. Degeneration of varietal purity after farmers
recycle seed for some time has been shown to be associated with a reduction in yielding
ability and productive potential of such recycled seed (Fehr, 1987). It is thus a
generalized recommendation that farmers should regularly use new, certified seed of
open pollinated varieties at least every third year (Diakite and Diarra 2000). Research
upon which such recommendations are based has been primarily conducted within the
context of commercialized farming systems of Europe or North America, where farmers
themselves normally do not invest their own efforts in selecting plants or grains
specifically for use as seed (Brick, 2004).

As highlighted by recent reviews, such as Jarvis et al., (2003), and by the seminal
work by Dr. Vavilov on centers of origin (Harlan, 1971), farmers play a crucial role in

conserving crop genetic variation and protecting the world’s agricultural heritage.

70

Farmers observe and select crops and crop varietiess, saving and managing seeds for the

next season. Thus, the traditional seed system has been the foundation of conservation of
biodiversity, and sources of resistance to abiotic and biotic stresses (Almekinders and de

Boef, 1999; Jarvis and Campilan, 2006).

Biodiversity and farmer’s selection criteria

Farmers maintain a set of crop populations that they keep as separate groups,
which are generally referred to as ‘varieties’ (Bellon, 1996). By maintaining and
continuing to select for varieties with specific sets of traits", farmers influence the
conservation of genetic diversity7 around the world, and the degree of biodiversity on
their farm (Jarvis, et al., 2008).

Using grains from their own harvest as seed (seed recycling), often combined
with selection of individual plants, inflorescences or grains are some of the means by
which farmers maintain seed stocks and thus genetic variation. The heterogeneity and
unpredictability of the environment that smallholder farmers face, the limited resources
which they can draw upon to manage shocks from pests and drought, are some of the
reasons that farmer’s maintain diversity (Rice et al., 1998). In addition, farmers select for
grain quality traits associated with multiple uses of crop products, including but not
limited to grain cooking characteristics, color, size and storability (Christinck et al.,

2005)

 

5 Variety: a plant grouping which is distinct in one or more forms or functions fi'om other such groups of a
plant of the same species and which maintains these distinctions when reproduced (Almekinder and
Louwaars, 1999).

6 A trait is a distinct variant of a phenotypic character of an organism that may be inherited.
environmentally determined or somewhere in between (http://en.wikipedia.org/wiki/Trait_(biological))

7 Genetic diversity is a level of biodiversity that refers to the total number of genetic characteristics in the
genetic makeup of a species (http://en.wikipedia.org/wiki/Genetic_diversity)

7l

Seed industry requirements for uniformity

Uniformity of crop varieties is usually achieved by genetic homogeneity of the
genotypes constituting the varieties, so that varieties will have few off-type8 plants and
the plants are uniform for plant height, morphological traits, such as glume color,
presence of leaf hairs and flowering date (Almekinders et Louwaars, 1999). Genetic
homogeneity is an important factor in obtaining pure stands of a specific variety (Brick,
2004). Variety uniformity is important when crops are produced for specific markets or
uses, e.g., for malting, sorghum grains need to be of uniform grain size, and have similar
germination times (Almekinders and Louwaars, 1999). Regulations for variety
registration and seed certification require specific levels of uniformity for each crop,

based on the biology of the crop.

In formal seed production processes, variety purity is crucial for seed lot
certification. It is assessed by field visits of the certification agents during critical stages
of crop development, usually just before flowering, especially for allogamous crops, and
just before harvest (Copeland and McDonald, 2001). Since varietal purity usually cannot
be determined by looking at the seed, seed certification programs and many seed
companies rely heavily on field inspection during seed production. Fingerprinting with
molecular markers is also used by practitioners to assure genetic homogeneity of seed
lots. Field inspection activities include: observation of plants’ agronomic performance,
the phenotypic uniformity (plant height, number of off-type plants), and observation of

flowering date. The laboratory test for variety uniformity generally includes observation

 

8 Off-type plant refers to a plant differing from the variety in morphological or other traits as a result of
mutation or cross-pollination

72

of grain uniformity through weight and size assessment, and grain vitrosity (Copeland

and McDonald, 2001).

Sorghum

Sorghum is used here as a case study to understand the impact of farmer seed saving
practices on variety purity. Sorghum is an important dietary staple crop in West Africa,
particularly in southern Mali (Koulikoro region) (Yapi et al., 2000). In Mali, sorghum is a
subsistence crop for most of the population, and the area it occupies is higher than that of
rice and maize (FAOSTAT 2007). Despite its importance to Malian food security,
sorghum faces some obstacles that are restraining its productivity. The average yield on-
farm in sorghum in Mali from 2000 to 2007 is about 1000 kg/ha (FAOSTAT 2008)
compared to 1.3t/ha in Ethiopia (Mekbib, 2005) and the 1.12t/ha average yield in
developing countries (FAOSTAT, 2005).

Sorghum (Sorghum bicolor) is a diploid (2n= 20), cereal crop and largely self-
pollinated (2-20% outcrossingg) (Minot et al., 2006, Almekinders and Louwaars 1999).
Based on this level of out-crossing, sorghum varieties are expected to be rather uniform
compared to millet (75 to 80% outcrossing) (Minot et al., 2006) and maize, which are
cross-pollinated crops (Almekinders and Louwaars 1999).

In Mali, local sorghum varieties collected from farmer sources have shown variability
among varieties for some traits, particularly for grain yield components in a field study
conducted by Chakanda (2000). This study was conducted in Mande zone Koulikoro
region of Mali. It was observed that a rather low level of variation occurred within farmer

varieties for traits studied such as date of flowering and panicle weight. Chakanda

 

9 Outcross or cross-pollinate occurs when pollen is delivered to a flower of a different plant under natural or
artificial conditions (http://en.wikipedia.org/wiki/Cross-pollinated#Mechanics)

73

concluded that these variations were probably caused by the way farmers manage their
fields and varieties. Almekinders (2001) as well as vom Brocke et a1. 2003, found that
intra-varietal diversity of farmers’ local varieties usually has a purpose. Sometimes
farmers use intra-varietal diversity to assure yield stability in highly unpredictable
production environments or to cope with variable or unpredictable pest and disease
dynamics (Diakite, 2003; ICRISAT, 2004; Almekinders, 2001). For sorghum in West

Afiica, such studies are rare.

Sorghum seed production by farmers in West Africa

Seed recycling is common among smallholder farmers in West Afiica
(Almekinders and Louwaars, I999). Diakité’s (2003) and Siart’s (2008) research on local
seed systems in Mali documented that seed recycling is a common practice among
farmers. From Diakité’s results, almost 50% of farmers in Mande zone have changed
(replaced) the sorghum varieties they grow only once, and about 16% never changed
varieties. The results of the same research in the Dioila zone showed that 35% of the
farmers interviewed changed their dominant sorghum varieties only one time and 6%
never changed theirs.

With regards to modern varietiesm, recycling is also common, at least among
Malian farmers. Siart (2008) reported from her field survey that more than 30% of
modern varieties grown in Dioila zone of Mali were recycled over 6-10 years and more
than 50% were recycled for 1-5 years. In this farm-level survey by Siart, in the Mande
zone similar if slightly lower levels of recycling were observed: 25-3 0% (6-10 years

recycling) and 30- 45% (1-5 years recycling).

 

'0 Modern variety is a variety developed by trained breeders working through targeted generation of
diversity, through crossing or other bio-technology tools, and selection (Almelinders and Louwaars, 1999)

74

Objectives

The objective of this research is to examine whether the seed selection practices used
by farmers in southern Mali for sorghum prevent degeneration, i.e. maintain the
productivity of the variety and maintain varietal purity, and thus intra—varietal
homogeneity. In this study, we use the opportunity to quantify genetic purity of seed lots
for four sorghum varieties that have been “recycled” for different amounts of time by
farmers who have been participating in variety selection research over several years and
have contributed to the identification of these varieties. The four sorghum varieties
(SAKOYKABA, SUMALEMBA, SUMBA, and TIEBLE) were developed jointly by
[ER and the ICRISAT/CIRAD sorghum breeding groups and have been in seed

production with farmers since 2003.

75

MATERIALS and METHODS

Research Activities

The field activities for this part of the research included:

0 Seed sample collection in the villages: 194 seed samples of four modern sorghum
varieties: TIEBLE, SUMALEMBA, SUMBA, and SAKOYKABA

0 Field trials with four modern ICRISAT sorghum varieties in production at the
farmers’ level

Table 2-1 : Places and approximate dates of the different activities undertaken to conduct
the field trial assessing the impact of farmers’ seed recycling on sorghum variety trait
uniformity from May to December 2007 in ICRISAT-Mali research station

 

 

 

Places Activities Time frame

16 villages including Seed sample
Dioila Seribila collection End of May, 2007
Zone Mangnambougou

Wakoro Wobougou Farmers survey

14 villages including Seed collection
Mande Gonsolo, Keniero End of May, 2007
zone Siranikoro Siby Farmers survey
ICRISAT research station Seed quality Field Trial From July to December,
(Samako experiment fields) 2007

Surveys

Seed sample collection

In May 2007, the seed samples were collected from farmers. At the same time, we
conducted a short survey. The village, the varieties and the number of seeds to be
collected per zone were determined in collaboration with ICRISAT researchers according
to their records of a survey conducted by village farmer facilitators during the 2005/6 dry
season. They had recorded in their specific villages who was growing which improved
variety since when (2003, 2004, or 2005) (Christinck and Weltzien, unpublished data).

Seed samples from 33 villages in the two zones were collected. The samples were from

76

farmers who had received one of the varieties for a trial or who had purchased seeds of at
least one the four varieties chosen for this trial. The seed collection was done along with
a short survey to get information about the seeds lot, such as first year of planting, seed
source, seed production and storage. The survey instrument is in Appendix 4.

The sample size of the seed sample collection was a function of the number of
seed lots of the different recycling categories we were studying. In Dioila zone, seed
samples were collected from 48 farmers in 16 villages, while in Mande zone there were
25 farmers in 14 villages. In the whole sample, there were only 9 women (7 in Dioila

zone and 2 in Mande). Almost all the farmers targeted (100%) were found.

Farmers ’ survey

During the general survey conducted at the farmers’ level for seed quality
assessment in December 2007, there was a part focused on farmers’ preferred dominant
varieties. This part of the survey aimed to assess, in addition to farmers’ preferences for
sorghum varieties, the genetic diversity at their level, i.e., the number of sorghum
varieties they were planting. The data collected on genetic diversity included the number
of varieties, the number of years of growing the same variety (recycling the seed), the
number of times farmers changed their preferred variety and the reason for this change.

The details of the survey i.e. the sample size, the villages where it was conducted

etc., are presented in Chapter 1 in the “Methods” section.

Assessment Varietal Purity or Homogeneity: Field Trial

A field trial was conducted at the ICRISAT research station with a range of

different seed lots of four sorghum varieties: TIEBLE, SUMALEMBA, SUMBA, and

77

SAKOYKABA, all developed through Participatory Plant Breeding (PPB) program and

in seed production at farmers’ level since 2003.

Experimental design

The split-plot design was used in the field trial, where the mean treatments were
the varieties and the sub-treatments were the seed lots (recycling), in the three
replications. The number of entries was 64, 16 seed lots for each variety X 4 (number of
varieties in trial) = 64 entries per block. These were replicated three times for a total of
64*3 = 192 plots. One hundred ninety-two (192) seeds lots split into three recycling
(foundation seeds, seed held by farmers for less than 3 years and seed held by farmers for
more than 3 years) were used .

Each plot was composed of 6 rows. The length of each row was 3 m, the space
between rows was 75 cm, and the space between sowing hills was 30 cm. On each row,
we had 11 sowing hills. The field was sown on July16‘ 2007. The field was thinned to

two plants per sowing hill on August 1, 2007, and the space between blocks was 1 m.

78

Material

Table 2-2: Description of the four sorghum varieties used for the field trial from the
catalog of sorghum varieties in Mali obtained with ICRISAT research technician

 

 

SAKOYKABA SUMBA SUMALEMBA Tieble
Origin Katibougou CEM Cameroun from the CSM 335
near Samanko 326/ 11 collection IS 15401 from Malian
CIRAD ICRISAT/CIRAD sorghum
collection
ICRISAT
/IER
Botanic Guinea, Caudatum, Guinea, late Guinea
Race, 'medium short flowering, erect medium
flowering duration, loose, duration, panicle cycle loose,
group drooping erect panicle drooping
panicle panicle
Plant 350- 400 150 - 250 440- 500 460
Height
(cm)
Plant Anthocyanee tan Anthocyanee Anthocyane
Color - e
Panicle Lax Elliptic Elliptic se‘mi- Lax
Shape semi- compact
compact
Panicle 42 - - 37 cm
length
(cm)
Photosens Photosensitive Not Very photosensitive Photosensiti
itivity photosensiti ve
ve
Growing 90 days 90 to 100 l 15 days 90 days
cycle days
Grain Big and tan Large grain size Big and
translucent White translucent
Rain-fed Yes Yes Yes Yes
crop
Preferred 800- 1000 mm 700-900 900 to 1200mm 800 to
rainfall mm 1 000mm
Average 2000-2500 kg 2000 — 2000 — 2300kg 1800 kg
yield/ha 2300kg
of grain
Cooking Very good in To Not good - Very good
characteri for To in To and
stics other local
dishes

 

79

We focused on seeds of improved varieties in production at the farmers’ level
since 2003: SAKOYKABA, SUMALEMBA, SUMBA, and TIEB LE. We tested 47 seed
lots from farmers produced without supervision, 13 seed lots from supervised seed
producers, and 4 breeder seed lots or foundation seeds (FS). Among the 47 seed lots
produced by farmers without supervision, there were four sub—lots composed of ES
(ICRISAT on-farm trial on 2003 seed), FSSP (farmers supervised seeds purchased) 2006,
FSSP 2005, and FSSP 2004. The 12 seed lots from farmer-supervised seeds (FSS)
include F88 2006. For analysis purposes, I grouped all seeds into three groups: FS, seed
held by farmers for less than 3 years (SHF<3), and seed held by farmers for more than 3
years (SHF>3).

Table 2-3: Number of seed samples (recycling categories and seed categories according

to the source) used in the field for each of the four varieties. Field trial conducted on
ICRISAT-Mali research station on 2007

 

 

 

Recycling Seed lots Number of seed lot collected per variety used in the
categories collected trial

Sakaykaba Sumalemba Sumba Tieble

Category 1 FS 1 l l 1

F SS 06 4 2 1 3

Category 2 FSSP 06 3 3 3 3

FSSP 05 l 3 2 3

Category 3 FSSP 04 4 3 4 4

ES 03 3 4 5 2

Field Monitoring

The emergence rate was evaluated by counting the number of empty sowing hills
seven days after planting. A score of 0 to 5 was given to each plot, 0 meaning excellent
germination and 5= poor germination. The seedling vigor was evaluated on the 15th day
after sowing by observing the height, width, and number of leaves on the plants, using the

same score. Thirty days after planting, the early vigor or plant growth was evaluated

80

through the height, width, and number of leaves on the plants. For both seedling and
development vigor, a score of l to 5 was given to each plot, 5 meaning excellent and 1:
poor

Phenotypic uniformity of each plot was evaluated by counting the number of off-
type plants in each plot, after flowering, just before harvest. A plot with no off-type
plants was considered to be uniform phenotypically and a plot with more than 5 off-type
plants was not uniform.

The heading date = the day the panicle emerges from the sheath, was recorded
very carefully by observing the date when the first 5 plants headed per plot, the 20%, the
50%, the 80% and the last 5 plants. The duration of the heading period was calculated as
the difference between the dates of the first 5% plants to the last 5% plants headed. The
height of 10 plants randomly chosen from the four central rows of the plot was measured
in centimeters 15 days after complete flowering. At harvest, the number of sowing hills to
be harvested was counted, and the number of panicles harvested was counted after
harvest. All six rows were harvested. The panicles were weighed before threshing, and

the grain was weighed immediately after threshing.

Data documentation and analysis
Surveys

Data were first registered on questionnaires for the seed sample collection surveys
in villages, in a field book for the field trial. We subsequently entered the data into a
Microsoft Excel spreadsheet. Excel was used to generate the frequencies and descriptive

statistics (means, standard deviation).

8]

Field trial

For the analysis of variance and Least Square Means (LSM) comparisons, I used
the Linear Mixed Models (PROC GLIMMIX) components of SAS. I had two fixed
factors (Varieties and recycling) in a split plot design, with the whole plot factor being
varieties (SAKOYKABA, SUMALEMBA, SUMBA and TIEBLE) in RCB (Randomly
Complete Block) fashion. The sub-plot factor was the seed recycling category or level of
recycling (foundation seed= 0, seed held by farmers for less three years <3, and seed held
by farmers for more than three years >3). We checked the homogeneity and normality of
the residuals of varieties and recycling categories using PROC Univariate and we found
tremendous heterogeneity of residuals in both varieties and recycling categories. Thus,
we decided to use heterogeneous GLIMM 1X models for the varieties and recycling.

To assess the effect of recycling categories within a given variety, we compare
Least square Means (LSM) for the interaction between variety and recycling categories
(variety*recycling).

Table 2-4: Description of the GLIMMEX model the statistical model used to analyze
the field trial data

 

Class information
Class Levels Values
Replication 3 1 5‘

 

Variety number 4 1- Sakoykaba
2- Sumalemba
3- Sumba
4- Tieblen
Category number 3 1- Foundation seeds
2- Seed held by farmers for more than 3 years
3- Seed held by farmers less than 3 years
Number of observation read 192
Number of observation used 192

 

82

RESULTS

Results from surveys

For this chapter, only the data from the farmers’ survey on varietal diversity for
sorghum were used for the analyses. Table 2-5 shows the number of seed samples
collected, the number of villages where the samples were collected, and the gender of the

growers surveyed by zone.

Table 2-5: Number of seed samples collected from farmers in Dioila and Mande zone for
the field trial conducted on the ICRISAT research station from July to December 2007 in
Mali

 

 

Zones #of seed growers # villages # women Varieties
Dioila 48 (66%) 16 7 (15%) All four varieties*
Mande 25 (34%) 14 2 (8%) All except SUMBA
Total 73 30 9 -

 

*The four varieties in the field trail were: SAKOYKABA, SUMALEMBA, SUMBA and TIEBLE

From the seed sample collection exercise, we documented a difference between
the two zones in terms of participation in research (twice as many farmers in Dioila than
in Mande) (Table 2-5). Sixty-six percent of the seed samples collected were from Dioila
and 34% from Mande. This is because Dioila zone is much bigger than Mande.

Another important finding was that in the whole sample there were only 9 women
(7 in Dioila zone and 2 in Mande). This reflects the weak involvement of women in
farmer participatory sorghum research in both zones and the fact that sorghum is not a

crop that women are responsible for in this region.

83

Table 2-6: Number of sorghum varieties grown by farmers from farmers’ survey
conducted on December 2007

 

 

Zto 3 varieties Two varieties One variety

(%) (%) (%)
Dioila n= 79 8 38 54
Mande n=75 8 52 40

 

The information in table 2-6 was needed to document the varietal diversity at
farmers’ level in the study zones. Table 2-6 shows that 8% of farmers in each zone grew
at least 3 varieties. However, a higher proportion of farmers in Mande (52%) grew two
varieties compared to Dioila (3 8%). More farmers in Dioila (54%) planted only one
sorghum variety (either local or improved variety) in their fields in 2007 than did farmers

in Mande (40%).

Field trial

The results from the field trial were grouped into three types of characteristics: the
heading (beginning, end and duration), the morphology of the plants (number of off-types

and plant height) and yield estimates (weight of harvested panicles and weight of the

grains).

Test of uniformity based on flowering

Table 2-7 presents results for different characteristics of flowering: the date of
heading of the first 5 plants to indicate the beginning of flowering (Ep5), the end of

heading (Epfin) and duration of heading (TEP).

84

Table 2-7: Results of GLIMMIX Procedure for flowering characteristics of the field trial
data, including the fixed effects of varieties, recycling categories and recycling categories
within variety and LSM comparisons among varieties and recycling categories, presented
as Julian days

 

 

Ep5 Epfin TEP
Response variables F P- F P- F P-
values values values
Varieties 84.25 <0.000 61.2 <0.000 0.9 0.46
l 1
Recycling 1.41 0.257 0.47 0.629 1.42 0.24
Variety* Recycling 1.23 0.46 1.47 0.29 1.82 0.10
Mean letter Mean letter Mean letter
(SE) (SE) (SE)
Variety SAKOYKABA 273 (1.7) b 281 (1.6) b 8.3 (1.2) a
SUMALEMBA 286 (1.0) a 293 (1.3) a 6.7 (1.0) a
SUMBA 268 (0.9) c 274 (1.3) c 6.8 (0.8) a
TIEBLE 271 (1.1) b 279 (1.5) b 8.1 (0.9) a
Means letter Means letter Means letter
(SE) (SE) (SE)
Recyclin Foundation seed 274 (1.3) a 281 (1.3) a 7.2 (1 .0) a
g =1
category SHF <3= 2 275 (0.7) a 282 (1.0) a 7.2 (0.6) a
SHF E =3 274(0.7) a 281(1 .0) a 7.9(0.6) a

 

The letters indicating the difference among varieties or recycling categories are read in this table by
column. Varieties and recycling categories with the same letter are not statistically at 95% significance
level different from each other.

The results from analyses of variance (table2-7) showed that variety had a
significant effect on Ep5 and Epfin levels (P-value <0.0001), but that recycling category
had no significant impact on any of the flowering variables. For TEP, only the interaction
variety *recycling category presented a mild significance (P-value= 0.10).

The LSM comparison results show that varieties SAKOYKABA and TIEBLE are
not significantly different in terms of beginning and end of flowering. These results agree
with the researcher’s assumption of variety description presented in table 2-1 that
SAKOYKABA and TIEBLE have the same growing cycle. The results of the comparison

among recycling categories within varieties are presented in table 2-8.

85

Table 2-8: Results of LSM comparisons of recycling categories within varieties for
flowering response variables (means, with standard errors in parentheses)

 

 

Varieties Recycling 1 Recycling 2 Recycling 3
Mean letter Mean letter Mean letter
(SE) (SE) (SE)
SAKOYKABA 274 (4.1) a 271 (1.7) a 271 (1.7) a
Ep5 SUMALEMBA 286 (1.7) a 286 (1) a 286 (1) a
SUMBA 268 (1 .4) b 268 (0.9) b 268 (0.9) b
TIEBLE 269 (2.2) b 271 (1.1) b 271 (1.1) b
SAKOYKABA 281(3.1) a 281 (1.6) a 281 (1.6) a
Epfin SUMALEMBA 293 (1.7) a 294 (1.3) a 294 (1.3) a
SUMBA 274 (1.6) a 275 (1 .2) a 275 (1.2) a
TIEBLE 277 (2.6) a 279 (1.4) a 279 (1.4) a
SAKOYKABA 7 (2.4) b 9.9 (1.2) a 9.9 (1.2) a
TEP SUMALEMBA 7 (1 .9) b 7.5 (1.0) b 7.5 (1.0) b
SUMBA 6.7 (1.2) b 6.6 (0.8) b 6.6 (0.8) b
TIEBLE 8.3 (1.6) a 7.7 (0.9) b 7.7 (0.9) b

 

The letters indicating the difference among varieties or recycling categories are read in this table by row.
Recycling categories with the same letter are not statistically different from each other.

Table 2-8 shows that researcher seeds (recycling category 1) have more variation
in the beginning and end of flowering than do farmer seeds. This variation is higher with
SAKOYKABA recycling 1 than the other three varieties. The results for Ep5 and Epfin
show that these variables did not differ significantly among recycling categories within
varieties, except for TEP of Tieble.

And for TEP, the results show that recycling 1 of SAKOYKABA was
significantly different from recycling 2 and 3. The flowering time for recycling 2 and 3
(=10 days) was long compared to recycling 1 (=7). For variety TIEBLE, recycling 1 was
significantly different from recycling 2 and 3. Recycling 1 had a longer flowering
duration (8.3 days) compared to recycling 2 and 3 (7.7 days), but only half a day, which

in fact has no practical relevance.

86

Test of uniformity based on plants morphology (appearances)

For this part, the analyses were done using the following two response variables:

plant height and the number of off-type plants in the plot.

Table 2-9: Results of GLIMMIX Procedure for plant phenotypic traits (plant height (cm)
and number of off-type plants for varieties, recycling categories alone and recycling
categories within variety and LS mean comparisons among varieties and recycling

 

 

categories
Plant height in centimeters #of off-types
(cm) plants/plot*
Effect F-values P-values F-values P-value
Varieties 42.61 <0.0001 3.24 0.04
Recycling 1.32 0.26 2.06 0.13
Variety* Recycling 0.73 0.85 1.12 0.42
Mean (SE) letter Mean (SE) letter
SAKOYKABA=1 391 (1 8.7) a 0.38(0.2) c
Varieti SUMALEMBA=2 403(8.1) a 2.1 l(0.8) a
es SUMBA=3 233(13.9) b 2.53 (0.9) a
TIEBLE=4 388(9.7) a 0.77 (0.3 ) b
Means (SE) letter Means (SE) letter
Foundation 352 (15.6) a 1.5 (0.8) a
Recycl seed=]
ing SHF<3= 2 349 (7.2) a 0.99 (0.3) a
SHF 2: =3 361 (7.1) a 1.87 (0.3) a

 

*With the plot of 6 lines, there were 1 1 sowing hills on each line and the field was thinned to 2 plants per

sowing hill. Thus the number of plants per plot was 11*6*2= 132.

The letters indicating the difference among varieties or recycling categories are read in this table by
column. Varieties and recycling categories with the same letter are not different from each other.

Table 2-9 shows that the varieties effect for plant height and number of off-type

plants (P-value= 0.0001 & 0.04) was significant at the 95% confidence level. Recycling

had a marginally significant effect on the number of off-type plants (P=O. l 3) and none on

plant height. No interaction of varieties* recycling categories was significant.

LSM comparisons for plant height for varieties showed that variety 3 (SUMBA)

was significantly different from the other three varieties (table2-9). SUMBA, with

87

average height of 233 centimeters, was shorter compared to the three other varieties,
which averaged about 400 centimeters in height.

The LSM comparisons for number of off-type plants showed that varieties
SUMBA and SUMALEMBA had more off-type plants (average 2 plants/plot 21.5%)
compared to TIEBLE (about 1 plant/plot 21%) and SAKOYKABA (almost zero off-type
plants observed).

Table 2-10: Results of LSM comparisons of recycling categories within varieties for plant

phenotypic response variables (plant height and number of off-type plants) (Means, with
standard errors in parentheses).

 

 

Resp. Varieties Recycling 1 Recycling 2 Recycling 3
Var Mean letter Mean letter Mean letter
(SE) (SE) (SE)
SAKOYKABA 384 (46.5) a 388 (17.7) a 402 (18.8) a
Plant SUMALEMBA 415 (12.7) a 396 (8.0) b 397 (8.2) b
height SUMBA 217 (32.4) b 227 (14.8) b 254 (12.7) b
(cm) TIEBLE 391 (19) a 382 (9.7) a 390 (9.5) a
# of SAKOYKABA 0.33 (0.6) a 0.29 (0.2) a 0.52 (0.2) a
0&1 SUMALEMBA 1.00 (2.2) c 1.50 (0.8) b 3.85 (0.8) a
type SUMBA 3.00 (2.3) a 1.94 (0.9) a 2.66 (0.8) a
plants TIEBLE 1.66 ( 0.7) a 0.23 (0.3) b 0.41 (0.2) b

 

The letters indicating the difference among varieties or recycling categories are read in this table by row.
Varieties within recycling categories with the same letter are not different from each other.

Table 2-10 shows that plant height for SAKOYKABA and SUMBA increases
with the duration of recycling. Here also the SEs for recycling category 1 are higher than
for recycling 2 and 3, with the higher SEs observed with SAKOYKABA and SUMBA
for plant height and SUMALEMBA and SUMBA for number of off-type plants.

The results of LSM comparisons of the effects of recycling within varieties on the
number of off-type plants revealed that recycling 3 of SUMALEMBA was significantly
different from recycling 1 and 2. Seed held by farmers for more than 3 years had more

off-type plants (average 3.8 plants/plot 23%) compared to researchers’ seeds (1.2

88

plants/plot =l%) and seeds held by farmers for less than 3 years (1.5 plants/plot =1 .1%).
These results agree with the research hypothesis 2 that varietal uniformity declined with
the duration of recycling.

For the variety TIEB LE, the results were the opposite of the research hypotheses
because recycling category 1 (researchers’ seeds) had more off-type plants than farmers’
seeds. This is interesting because it shows that farmers are reducing the heterogeneity

within the variety beyond that of the foundation seed.

Test of uniformity based on yield estimates

Two response variables, weight of harvested panicles (normally called panicle
yield) and the weight of their grains, grain yield, were used for this part. Table 2-9
presents the results of the LSM analyses.
Table 2-11: Results of GLIMMIX Procedure for plant yield components (panicle yield
and grain yield (fixed effects) of varieties, recycling categories alone and recycling

categories within variety and LS mean comparisons among varieties and recycling
categories

 

 

Weight of harvested Weight of ains
panicles (g/ m )
(g/ m”
Effect F-values P-values F -values P-value
Varieties 3.83 0.046 4.09 0.036
Recycling 2.23 0.1 1 2.72 0.069
Variety* Recycling 2.42 0.037 2.3 1 0.042
Mean (SE) letter Mean (SE) letter
SAKOYKABA 346 (27.2) a 260 (20.8) a
Variety SUMALEMBA 267 (25.7) b 198 (19.2) b
SUMBA 299 (26.3) c 238 (20.0) c
TIEBLE 360 (30.6) a 280 (23.6) a
Means (SE) letter Means (SE) letter
Foundation 340 (26.8) a 263 (20.6) a
Recycli seed=1
ng SHF <3= 2 312 (19.8) a 241 (14.4) a
SHF 23 =3 300 (19.7) a 228 (14.4) a

 

The letters indicating the difference among varieties or recycling categories are read in this table by
column. Varieties and recycling categories with the same letter are not different from each other.

89

The results of the analyses of variance showed that varieties in general and the
recycling categories for specific varieties differed significantly for both the weight of
harvested panicles and grains at the 95% level of confidence. For recycling categories,
there was trend towards lower yields as recycling increased, at an 80% level of
significance. The weight of panicles for TIEBLE and SAKOYKABA were not
significantly different from one another. Their harvested panicles weighed on average
about 346 and 360 g/m2. They were significantly different from SUMBA and
SUMALEMBA, whose harvested panicles weighed respectively 299 and 267 g/ m.2
Based on these results, we conclude that SUMALEMBA panicles weighed less than
those of the other three varieties.

In terms of grain yield, all four varieties had about 2 t/ha. The results for grain
weight were similar to panicle weight. TIEBLE and SAKOYKABA had consistently
higher yields than did varieties SUMBA and SUMALEMBA (Table 2-9).

The results of the interaction of varieties*recycling for harvested panicles and

grains weight are shown in table 2-12.

90

Table 2-12: Results of LSM comparisons of recycling categories within varieties for plant
yield response variables (panicle yield and grain yield) (Means, with standard errors in
parentheses)

 

 

Resp. Varieties Recycling 1 Recycling 2 Recycling 3
Var Mean lett Mean letter Mean letter
(SE) er (SE) (SE)
Weight SAKOYKABA 348 b 374 (26.7) a 315 c
of (44.2) (27.2)
Panicles SUMALEMBA 294 a 271 (25.4) a 236 b
(g/ m2) (37.3) (25.7)
SUMBA 305 a 302 (26.7) a 289 a
(40.0) (25.6)
TIEBLE 415 a 303 (30.7) c 360 b
(57.9) (29.8)
Grains SAKOYKABA 262 b 285(20.4) a 234(20. c
Weight (35.4) 8)
(g/ m2 ) SUMALEMBA 219 a 202 (19.0) a 173 b
(28.6) (19.3)
SUMBA 244 a 242 (20.4) a 229 a
(31.9) (19.4)
TIEBLE 328 b 236(23.7) a 276 c
(46.4) (23.0)

 

The letters indicating the difference among varieties or recycling categories are read in this table by row.

Varieties and recycling categories with the same letter are not different from each other.

From table 2-12, the weight of harvested panicles of recycling groups 1, 2 and 3

of SAKOYKABA are all different from each other, with recycling 2 category weighing

more (374g/ m2) than the two others. For SUMALEMBA recycling 1 and 2 are not

significantly different from each other, but they are different from recycling 3. The

panicles of SUMALEMBA recycling 1 and 2 weighed 294 and 271 g/ mz‘ respectively,

while recycling 3 panicles weighed only 236 g/ m2. For variety SUMBA, there is no

significant difference among the three recycling categories in terms of the weight of

harvested panicles. For variety TIEBLE, the harvested panicles of recycling category I

weighed more (415 g/ m2) than that of recycling category 3 (360 g/ m2), which also

weighed more than that of recycling category 2 (303g/ m2).

91

Yield of grains showed an overall pattern of response similar to that of panicle

weight.

Farmers ’ surveys

The farmers’ survey documented how long farmers grew the same variety, and if
they happened to change varieties, how many times they changed and why (table 2-13).

Table 2-13: Farmers genetic diversity management for sorghum assessed by recording
the duration of farmer growing the same variety and the percentage of farmers who
reported of having change their preferred dominant variety at least one time during
farmer general surveys conducted in Dioila and Mande zones

 

 

Dioila n=79 Mande n=75

Duration of growing (# of years) Mean SE Mean SE

10 (1.1) 10.3 (1.2)
Change of the preferred variety (%) 8 17
Reason of change n*=6 n*=13
Decrease in yield 2
Rain irregularity 1 6
Soil fertility problem 2
No response 3 5

 

* n =number of farmers in survey sample who changed their preferred sorghum variety at least once in both
zones, where the research was conducted in December 2007.

As shown in table 2-13, farmers in both zones have been growing the same
dominant sorghum variety for 10 years on average. Seventeen percent of farmers in
Mande have changed their preferred sorghum variety at least one time against 8% in the
Dioila zone. In the Dioila zone, the most popular reason for change was No response
(50%), while in Mande the irregularity of rains was the most frequent reason for variety

change (6 of the 13 farmers who changed varieties).

92

DISCUSSION

Farmers’ diversity management

The analysis of farmers’ management of varietal diversity in this study showed
that farmers in both zones grew the same dominant preferred varieties for more than 10
years on average (table 2-13). This recycling duration is even longer than the one in
Siart’s (2005) survey results in the same zones.

In 2007, about 60% of farmers from Mande had at least two sorghum varieties in
their fields, compared with 46% in Dioila (table 2-6). This suggests that despite the
availability of research and extension services in Dioila, the farmers Dioila remain
relatively closed to sorghum variety diversity. Siart (2005 survey reported in 2008)
found that the majority of households in Dioila (56- 69%) and in Mande (57-62%) grew
more than one variety. From her findings, Dioila farmers had more varietal diversity than

Mande farmers. However, she noted that these numbers varied from year to year.

Variety Mixture or Misidentification by Farmers

The same varieties that are cultivated by different farmers over many years are
likely to lose their identity for certain traits (Chakanda, 2000). The data analyses from the
field trial revealed the influence of farmers’ seed management and variety saving on
modern sorghum varieties in production on-farm since 2003.

Chakanda (2000) also found in a Mali seed survey that despite the effectiveness
of annual selection of panicles, there was a high level of variation for traits within
farmers’ varieties. Most of the time, due to the visual resemblance in all other traits,
farmers believe that they have maintained purity of varieties and are growing a specific

variety. This was noticed in the field trial with the fake TIEBLE variety we had, as all the

93

other varietal traits were the same as TIEBLE except the grain appearance. There might

be both negative and positive aspects of “renaming” a new variety by farmers.

Field Uniformity Experiment

Even though sorghum is largely self pollinated, especially in guinea race varieties
we tend to get 20 sometime 25% outcrossing (Minot et al., 2006). This also means that
we get actually fairly high levels of heterozygosity in these varieties after some years of
recycling. One of the consequences of keeping too much homogeneity is that it may
entail inbreeding, and thus a reduction in yielding ability. For these sorghum varieties it is
thus important to maintain some level of diversity, for some traits. Actually the Tieble
results show this tendency: with more variability higher yield (recycling 1, than with
reduced variability, lees yield.

In the field trial, we found that although there was some variation in variety traits,
overall there was generally uniform expression of traits across different recycling
categories within varieties. Thus, despite occasional “renaming” of a variety such as seen
with TIEBLE case, overall farmers were selecting and maintaining pure varieties.

F lowering] I

The results of the uniformity assessment based on the timing of flowering showed
that no significant difference existed between researcher seeds (recycling 1) and farmers
recycled seeds (recycling 2 and 3). The flowering time of varieties SAKOYKABA and
TIEBLE were not significantly different from each other, but they were different from
SUMALEMBA and SUMBA. This is consistent with the descriptions of SUMBA as an

early maturing variety and of SUMALEMBA a late maturing variety (as its name

 

H This is the first study we are aware of that examined the impact of the duration of fanner-recycled seeds
versus researcher-produced seeds on uniformity of flowering traits.

94

indicates in the local language Bambara), while TIEBLE and SAKOYKABA are
intermediate cycle (table 2-1) (ref Mali National Seed Service variety catalog).

Examining the results in more detail, although we note no significant difference
among the interaction of recycling category*variety, a trend suggests that researcher
seeds (recyclingl) of TIEB LE flowered earlier compared to seed held by farmers
(recycling 2 and 3). And for Epfin, recycling] and seeds held by farmers for less than 3
years (recycling 2) were apparently different from seeds held by farmers for more than 3
years (recycling 3). Our research hypothesis was that recycling 1 seeds have more
uniform varietal traits compared to farmers’ seeds and that trait variation increases as
seeds are recycled more (Chakanda, 2000). For flowering traits, Ep5 and Epfin, variations
observed in variety TIEBLE may have been due to seed recycling.

When considering the duration of the flowering period (TEP), only the interaction
of varieties*recycling was significant (Table 2-7). The LSM analysis of
varieties*recycling showed that the recycling categories within SAKOYKABA were
different. For SAKOYKABA, researcher seed had the shortest flowering duration (7
days) (table 2-8). Thus, the TEP for SAKOYKABA increased with increasing time of
recycling by farmers. For TIEBLE, recycling 1 had longer flowering duration (8 days)
than farmers’ seeds. These results show that even researchers have some uniformity
problems with TIEBLE, and to some extent one can say that farmers did a better job
maintaining flowering uniformity (table 2-8).

The TEP conveys information about the uniformity of maturity and to some

extent variety uniformity. Thus, the shorter the flowering duration is, the more uniform is

95

the maturity, making the harvest easier for farmers because a longer period over which

maturity will occur will stretch out harvest duration (Rasabandit et al., 2006).

Morphology

The test of uniformity based on plants’ morphological traits (height and number
of off-types) showed an effect of variety on plant height. Varieties SUMALEMBA,
SAKOYKABA and TIEB LE were not different from each other, but they were all taller
than SUMBA. This result is consistent with the researcher description of SUMBA.

The standard error for plant height is consistently higher among researcher-
produced seeds (table 2-10). Also, the height of SUMBA increases with the recycling
duration (table 2-10). This result suggests that farmers may be selecting SUMBA for
taller height. As in many developing countries, in Mali, sorghum stover is often used to
feed cattle (de Vries and Toenniessen 2001 cited by F olkentsma et al., 2005), and farmers
plant a variety for multiple uses to meet their needs (Diakite, 2003).

The test for off-type plants showed that both varieties and recycling were
significant, although recycling was significant only at the 13% level of significance (table
2—9). The varieties varied markedly in terms of number of off-type plants. SUMBA and
SUMALEMBA had more off-type plants than TIEBLE and SAKOYKABA. This finding
is consistent with earlier reports by researchers. However, the good news is that the
number of off-types in all varieties was less than 5% (the critical number of off-types)
The seeds in the longest recycling category showed more off-types for SUMALEMBA
(table 2-10). However, with TIEBLE (excluding the “fake” TIEB LE), the trend was for
researcher seeds to have more off-types than farmers seeds. This result shows that

farmers are doing a better job in selecting TIEBLE seeds than are researchers.

96

Yield

The analyses of recycling effects on yield components showed that both varieties
and recycling had significant influence on grain and panicle weight. The varieties
TIEBLE and SAKOYKABA had the highest weight of harvested panicles and grains.
This is not consistent with the description of the varieties in table 2-2, where TIEBLE at
(1800kg/ha) was the lowest yielding of the four varieties. The yield observed in my field
trial was very high (example of TIEBLE 23000 kg/ha). This yield may be due to
favorable management and weather conditions of the trial.

The analyses of the interaction of varieties*recycling showed the overall yield
was not affected by recycling. Especially for TIEB LE, recycling 3 had higher panicle
and grain weight compared to recycling 2. The duration of recycling was not associated

with low yield.

97

CONCLUSION

The field trial analyses showed that there are no significant differences between
seeds of modern varieties produced by farmers without supervision and foundation seeds
in terms of flowering (beginning and end), number of off- types, and panicle weight. That
means that farmers are producing seeds with good level 8 of varietal purity, and that
varietal identify is well maintained even for varieties that individual farmers have
maintained for four consecutive years. Since this field trial was conducted on the research
station only, for more comparisons it will be good to have some trials at the farmers’
level in the future research.

In conclusion, farmers’ seed recycling does affect some varietal traits of sorghum,

but the effects are not serious enough to worry about them.

98

CHAPTER 3

SORGHUM VARIETY TRAIT EVALUATION WITH
FARMERS IN SOUTHERN MALI

INTRODUCTION

Sorghum (Sorghum bicolor (L) is the fifth most important cereal crop worldwide
(Mekbib, 2006) and the third most important cereal crop in Mali (Chakanda, 2000). Since
sorghum is a C4 plant, it shows a greater efficiency of dry matter production relative to
water use than rice and wheat (Downess, 1970). It can also endure short periods of
waterlogging (Dogget and Jowett, 1966) and has considerable tolerance to drought and
heat stress (Downess, 1970). This is why it is grown mainly in the drier regions of West
Afiica (Chakanda, 2000).

In Mali, sorghum is the staple food crop in the 700 - 1200 mmrainfall zones
(Almekinders et al., 2007) such as the Koulikoro region investigated in this study. It is
mainly a subsistence crop for most of the population in the Sudanaan zone of Mali (Byth,
1993), and the number one food crop in the two study zones (Dioila and Maude). In
addition to food production, sorghum is used in local beer production, and its stalks are
used for animal feed, for construction, or as fire wood (Mburu, 1986).

Despite its importance, sorghum productivity is still very low in Mali: the average
yield is 1.0 t/ha (FAOSTAT, 2008) compared to an average yield of 1.12 t/ha in
developing countries and a country-wide yield of 1.3t/ha in Ethiopia (Mekbib, 2005).
Researchers associate this low yield of sorghum with the very low levels of fertilizers

applied in general (8 kg/ha), and to sorghum in particular, and to the use of local varieties

99

(Diakite, 2003). In Mali, as well as many part of Africa, traditional varieties are often
preferred by farmers and have been shown to be well adapted to biotic and abiotic stress
factors and exhibit superior grain quality traits for the production of local dishes. These
are still the main varieties cultivated in southern Afiica (Shargie and Wenzel 2008) as
well as in Mali (Siart, 2008).

In the past, agricultural research in Mali has often focused on cash crops such as
cotton and rice (Tefft, 2004). Recent problems of food insecurity in Mali have
encouraged research institutions and the government to focus on agronomic performance
of staple grain crops (sorghum, millet, corn, etc.). Just after Mali’s independence, the
Institute of Rural Economic (IER) was created to find ways to improve the productivity
of food crops (Yapi et al., 2000). Malian researchers collaborate with sorghum breeders
from international institutions (such as ICRISAT), and have recently developed varieties
with high yield potential to increase food security for the majority of the Malian
population (Chakanda, 2000). Because of the importance of the seed sector in agricultural
productivity, researchers have tried to improve the seed sector by promoting
multiplication and distribution of improved variety seeds (Diakite, 2003, Maredia et al.,
1999). A number of improved sorghum varieties such as SUMBA, TIEBLE, and
SUMALEM BA have been developed and disseminated in the country through the
[ER/ICRISAT Participatory Plant Breeding program (PPB)

The formal seed sector in Mali does not function well for subsistence or modestly
valued crops such as a sorghum. This is common among SSA countries, which have
failed to improve farmer access to high quality seeds of widely grown subsistence crops.

There are many reasons for this, among them the high cost of seeds relative to cereal

100

grain prices, the limited number of appropriate varieties that meet the needs of farmers,
inconsistent seed quality, and the dispersed location of seed production. Farmers are
generally considered as consumers, rather than as seed producers and key participants in
seed assessment and dissemination (Chakanda, 2000, Maredia et al., 1999).

Participatory plant breeding is becoming a widely recognized approach to
improve client-orientation of research and develop more adoptable in varieties. ICRISAT
is pursing this strategy in Mali, using Farmer Participatory Variety Selection (F PVS)
(Weltzien et al., 2006, Vemooy 2003). This methods can include farmers at every step of
the breeding process (Friis-Hansen and Sthapit, 2000; Vemooy 2003). Farmer
participation in technology generation can enhance the effectiveness of selection and has
been show to improve adoption (Ashby 1990; Sperling et al., 1993; N’tare et al., 2008).
Farmer involvement in variety evaluation provides means for identifying a wide range of
locally valued varietal traits ( N’tare et al., 2008; Baidu-Forson, 1997, Almekinders et al.,
1999)

PVS is a useful research communication tool that facilitates researcher and farmer
interaction. It enhances the ability of plant breeders to quantify the performance of new
genotypes across the wide variability of growing conditions in a target area and helps to
assess the response of varieties to farmer management (Rasabandit et al., 2006).

Seed system analyses conducted in Mali have shown that farmers are interested in
trying new types of varieties that combine preferred variety traits (Weltzien et al., 2005).
An opportunity is presented by the shift over the last decade in plant breeding priorities in
Mali to incorporate farmer selection criteria. This provides a new environment, with the

development of farmer-selected sorghum varieties.

101

Research Objectives

This is a participatory research project, conducted in collaboration with farmers.
It aims to quantify farmers’ evaluation of variety traits in recently developed improved
sorghum varieties, and provide information on farmer preferences to be incorporated into
the plant breeding priorities of the ICRISAT and IER sorghum breeding programs in

Mali.

Hypotheses

We tested the following hypotheses:

0 Farmers in both study zones have the same preferred variety traits for sorghum.

0 There is no gender specificity in terms of variety trait preference.

0 We hypothesize that under irregular and low rainfall farming conditions, yield,
early maturity and drought resistance are the three most popular varietal traits in

the study area.

102

MATERIALS AND METHODS

A Participatory Variety Selection (PVS) method was used to assess farmers’
preferences for modern sorghum varieties in trials conducted in the Dioila and Mande
zones. The trials were planned jointly by farmers and researchers and managed by
farmers (Weltzien et al. 2007). Evaluation of the trials was done jointly, by farmers and
researchers, each evaluating specific traits. In addition, a survey was conducted to
document farmer assessment of all varieties they have experience with, including the

varieties they have seen in the on-farm trials and the varieties they are currently growing.

Participatory variety selection trials

The location of the trials and the dates of farming activities were chosen by
farmers, whereas inputs in the form of seed, fertilizer and technical support such as the
trial design were provided by plant breeders from ICRISAT and IER (Weltzien et al.,
2006).

The trials were conducted to evaluate 32 varieties (16 tall and 16 short varieties)
at on-farm sites located in 9 villages (five in Dioila and four in Mande). In each village,
two trials were established, each with four replications, i.e. twp replications per farmer.
One compared the short varieties and the other the tall varieties. Thus, 14 modern
varieties were compared, plus one local and one common control variety in each of these
trials. The local variety was one of the main varieties used in the village and was chosen
by four farmer participants (Weltzien et al., 2007). The common control variety for all
the trials was TIEBLE (CSM 335). The experiments were conducted in farmers’ main
sorghum production fields. The plots had 6 rows of 5 meter length each arranged in 4

ranges of eight plots and randomized as alpha lattice designs with four plots per block

103

(Weltzien et al., 2007). The trials were managed by farmers, with support from a farmer

organization secretary, whose position was called an “Animateur” and by visits by

ICRISAT and IER technicians. This supervision was an important means of supporting

farmers to be involved in evaluation, as not all farmers were literate.

Many farmers were involved in the variety evaluation conducted at harvest time

(end of November and beginning of December 2007), just before harvest when the grains

were mature and easy to observe.

Table 3-1: Name of the 32 varieties that farmers evaluated to assess farmers variety trait
preference in December 2007 in the Dioila and Mande zones

 

Farmer evaluation

Entry #

EGESSZSOWQOMAwN~

Variety names

Tall type
Sekounioni
Nionifing
Yebagasago
Bibagalawili
Bamouka
Kassoroka
Bandokablen
Dougouba
Sobani
Yoka
Palo
Yalama
Koroba
Niakafa
Yamasa
Tieble

Short type
Kakou
Boulen
Djelefi

Oki
Tamia
Koule
Lebo
Ti guila
Gagna
Kouladji
Drasa
Tiandougou Coura

Marakanio
Grinkan
Tieble

Local

There were two types of farmer evaluations carried out. First, a simple scoring

method, where farmers chose varieties in the experiment, was used to document farmers’

expressions’ of preference for the top three preferred varieties. This approach is called

‘voting for varieties’ (Christinck et 2005, Weltzien et a1. 2007). Second, a survey was

104

carried out to document farmers’ assessments of the varieties they grow. For the voting

exercise, farmers first conducted a visual evaluation of plants standing in the field, based

on the overall plant appearances (strength of the stem, appearance of the panicles, grains,
leaves). Farmers first scored all 16 varieties for ICRISAT/IER FPVS. Ballots of three

' different colored papers (white, yellow, and red) were used for voting.

In addition, farmers were asked to select the three best varieties using pieces of
cloth with different numbers of lines on each. Each farmer had to tie the piece with one
line at the beginning of the field plot that was his/her first or best, then the piece with two
lines on the second best, and the third on the third best. After this evaluation, the
technicians conducted a short survey about the reasons for the farmers’ choices.

These evaluations were conducted in 9 villages by me, working in close
collaboration with the plant breeder evaluation team from ICRISAT/IER. As done by
Rasabandit et al. (2006) for rice PVS, the evaluation was done directly in the field by

farmers.

Farmer appointment and field preparation

The fields (replication) where the evaluation was to take place were chosen jointly
by the research technicians and me, and the farmers conducting these trials were
informed ahead of time about the chosen locations. The animateurs of the villages were
informed about the dates of the evaluations, and they had to inform all the farmers (men
and women).

Early in the morning of the day of the evaluation, the research technicians and the
owner of the trial headed into the field to prepare it for voting. A paper bag was hung tied

to the stems of the first plants in a plot for receiving ballots from farmers during voting.

105

A shade tree was identified for gathering farmers for the quick survey on the reasons of

their choices.

Farmers ’ evaluation

Farmers were asked to express their preferences for individual varieties by
scoring them using colored papers. A white paper indicated that the farmer was interested
in growing this variety, yellow indicated the varieties that might require more testing and
red indicated varieties that the farmer rejected. Each farmer was given an envelope
containing about 30 pieces of paper of each color and the three pieces of cloth material
(Christinck et 2005, Weltzien et al. 2007).

Before making a decision, farmers were asked to carefully observe and consider
the performance of variety in each plot. The procedure was carefully explained to
farmers, beforehand and again in the field just before starting the evaluation. Thus, the
evaluation involved an initial field tour where all farmers walked around the field to look
at the different varieties. During the second round farmer sscored the varieties . During

the third round, the farrmers identified the three best varieties.

Samples

The goal was to involve equally women and men farmers; thus, I used a purposive
selection process. Upon arrival in the village, the first 30 or 40 people were selected.
However, when there were only men in the first 30 or 40, we postponed the selection
until about 50% or so of the farmers participating were women. In a village where the
number of farmers showing up was less than 30, everybody was selected. In total, 230
farmers (124 men and 106 women) participated in this part on the research. For the

Dioila zone, there were 141 farmers from Wobougou (trial hybrid), Kafla, Seribila,

106

Kegne, and Wakoro. For the Mande zone, there were 89 farmers from Teneya, Keniero,
Kalague, and Gonsolo.
Table 3-2: Number of farmers by gender who participated in the on-farm farmers’

sorghum evaluation exercise aimed to assess farmers’ variety preferences in the Dioila
and Mande zones in December, 2007

 

 

Men Women Total/zone
Dioila 71 70 141
Mande 53 36 89
Total/ gender 1 24 106 230

 

General farmers’ survey

This survey is described in chapter 1. The survey provided an opportunity to
document farmers’ preference for variety traits. For detailed information about the survey
process please refer to chapter 1. The survey was conducted from the end of November to
mid December 2007, in the Dioila and Mande zones of Mali and involved interviewing
157 farmers in 8 villages. Each farmer were asked about his/her preferred dominant
sorghum variety (name, source, village, and time of acquisition). I then documented the
traits that the farmer described as being the reasons this variety was preferred. The
variety traits in the survey included: yield, resistance to insects, early maturity, grain

cooking traits, drought tolerance, secondary uses (like animal feed), and other.

Data documentation and analyses

The data for this chapter were collected by me and through close supervision of
eight experienced assistants. These included four enumerators, two technicians from IER
and two from ICRISAT, all of whom had been trained in conducting farmer variety
evaluation surveys. It was important to conduct the evaluation in a timely fashion in order

to meet farmer time constraints, thus, a large group of enumerators was required.

107

To synthesize the data, I grouped the variety traits described by farmers from the
on-farm evaluation into six groups: yield, early maturity, resistance to drought and striga,
grain appearance, adaptation to the farming areas and others. The details regarding
farmer trait descriptions and how these were grouped into 6 categories are presented in
Appendix 9.

The data were first recorded on notebooks and questionnaires in the field. They
were then entered on a laptop computer in Microsoft Office Excel, which was used to

generate standard descriptive statistics and frequency analyses.

108

RESULTS

On-farm field trial evaluation by farmers

The results of farruers’ voting of varieties is presented in table 3-3. The trials for
tall and short varieties were conducted separately at different farmers’ sites to avoid that
the short varieties get shaded by the tall varieties. Thus, genotypes were evaluated by
farmers within each height category. As shown in table 3-3, there were notable
differences between the Dioila and Maude sites in terms of farmer preferences. Gender,
on the other hand, did not have a marked effect, at least in terms of farmer of the field
performance: Women and men from Dioila tended to favor entries 1, 2, 13 and 14 among
the tall varieties and 1, 14 and 15 among the short types (although women did not rank
any variety except #1 above 70% among short types). In Maude, several of the same tall
varieties were liked (I, 2, l4 and 16); however, short variety preference was quite

different than that observed in Dioila (Table 3-3).

109

Table 3-3: Farmers’ preferences expressed in percentage for the 32 varieties in the on-
farm trials by gender, variety type (short versus tall types) and by zone (Dioila and
Maude) from the ICRISAT/ IER in December 2007

 

 

Dioila (%) Maude (%)
Entry Tall type Short types Tall type Short type
# men women men women men women men women
1 81 80 72 72 83 78 53 50
2 87 81 38 45 86 81 34 57
3 45 46 28 35 79 69 53 53
4 46 49 29 39 80 73 38 52
5 66 73 28 33 71 71 46 51
6 60 58 23 46 60 63 40 52
7 56 68 23 35 58 60 41 47
8 51 52 58 51 51 60 51 59
9 47 52 35 40 49 57 22 35
10 18 32 30 30 27 36 32 47
11 22 39 64 58 23 35 61 60
12 34 46 65 66 54 60 56 58
13 76 71 28 30 63 71 35 36
14 76 73 70 68 68 68 53 57
15 67 61 82 65 51 56 92 84
16 68 52 63 63 76 75 86 74

 

Source: Results of ICRISAT/IER on-farru field trial evaluation, 2007.

Farmers were asked to next select their top three varieties and describe the traits
that led to that selection. The table 3-4 presents the results from this second step in the
farruers’ evaluation of sorghum varieties grown in on-farm field trials.

Table 3-4: The results of reasons why farrmers from Dioila and Maude zones (n=230)

rated the sorghum varieties as their three top choices from the on-farm field trials in
December 2007

 

 

Preferred varietal traits lst choice* 2nd choice 3rd choice
11 % n % n %
Yield 178 77 137 60 131 57
Early maturing 45 19 22 9 18 8
Drought and Striga tolerance 27 12 23 10 15 6
Grain appearance 129 56 99 43 95 41
Adaptation to the cultivation area 47 20 26 11 25 1 1
Others 11 5 8 3 6 3

 

The percentages summed are greater than 100% because farmers were allowed to cite as many variety traits
as many as they thought were important.

110

As shown in table 3-4, yield traits were noted as keys factor by 77% of farmers
for their top pick 77%, 60% of farmers for their second pick and 57% of farmers for their
third pick. Grain appearance was also important, with 56% of farmers noting this for
their fist pick, 43% for their second pick, and 41% for their third pick. The third most
preferred variety trait was adaptation to the field environment, which was noted by 20%
of farmers for their first choice, and by 11% for the second and third choices.

A more in-depth understanding of farmer preferred varietal traits was arrived at
by evaluating responses for the two locations, Dioila and Maude. As shown in table 3-5,
the variety traits assessed in on-farm trials at both locations prioritized yield and grain
appearance, but in addition early maturing varieties were valued by farmers particularly
in Dioila (18%), and resistance to drought and striga was noted by 8% of farmers at both
sites (table 3-5)

Table 3-5: Results of farmers’ variety traits evaluation done on December 2007 by zone

of the on-farm trials, expressed as the number of farmers and percentage of the sample
citing different reasons for their variety preferences, per zone in southern Mali

 

Dioila n=l41 Maude n=89

 

Varietal traits n % n %
Yield 93 66 55 62
Early maturity 25 18 4 4
Resistance to drought & Striga l2 8 7 8
Grain appearance 61 43 46 52
Adaptation to the area 28 20 9 10
Others 6 4 3 3

 

The percentages summed are greater than 100% because farmers were allowed to cite as many variety traits
as they thought were important.

The results in table 3-5 show that yield, listed by 66% of farmers in Dioila and
62% in Maude, was the most popular varietal trait that farmers in both zones were

looking for in new sorghum varieties. Yield is followed by grain appearance in second

111

place with 43% in the Dioila zone versus 52% in Maude. Then adaptation to the

cultivation area is the third most popular varietal for farmers in both zones.

Gender considerations in varietal trait evaluation

Table 3-6: Results of farmers’ variety traits evaluation done in December 2007, by
gender

 

 

Men n=124 Women n=106
Varietal traits n % n %
Yield 80 64 68 64
Early maturity 16 13 1 1 10
Resistance to drought & Striga 12 10 6 6
Grain appearance 55 44 51 48
Adaptation to the area 23 18 13 12
Others 6 5 2 2

 

The percentages summed are greater than 100% because farmers were allowed to cite as many variety traits
as they thought were important.

The results in table 3-6 show that varietal trait preference for new sorghum
varieties are the same for both men and women in the study areas. Yield was still the
most important varietal trait for both men and women (64%). This result agrees with our
research hypothesis.

Table 3-7: Results of farmers’ variety traits evaluation done in December 2007, by
gender and by zone

 

 

Dioila n= 141 Maude n= 89
Women Women

Varietal traits Men n=71 n=70 Men n=53 n= 36

% % % %
Yield 64 68 61 67
Early maturing 17 13 5 6
Resistance to drought & striga 7 9 10 6
Grain appearance 46 41 57 45
Adaptation to the area 20 20 15 4
Others 5 2 4 3

 

The percentages summed are greater than 100% because farmers were allowed to cite as many variety traits
as they thought were important.

112

The results by zone and by gender about farmers’ preferred varietal traits are the
same as for the entire sample, with the only exception being that women in Maude
preferred early maturity and resistance to drought and Striga rather than adaptation to the

cultivation area.

Farmer survey

Table 3-8: The name of dominant preferred sorghum varieties (varieties grown in large
space by farmers encountered in Dioila and Maude. The table also shows the frequency
(%) (the number of time a variety was cited as dominant by farmers). The data to make
this was from farmers’ general survey conducted in December 2007.

 

 

Order Dioila Maude
Variety Variety Frequency Variety name Variety Frequency
name type (%) type (%)
1 Nionbleui LCV 16.4 Tiemarifing LCV 22.7
2 Bandoka FV 1 5.2 Seguetana LCV 1 3 .3
07/08
3 Seguetana LCV 7.6 Keleyabomusola FV 8
07/08
4 Tieble MV 5 Doronkonikalan LCV 5.3
5 Yubleni LCV 5 Tieble MV 4
6 Gnegnebleni LCV 5 Niagafing FV 4
7 Kassaroka FV 5 Kalosabani LCV 4
07/08
8 Fambe MV 2.5 Touroukani LCV 2.7
9 Sumba MV 2.5 Kenike LCV 2.7
10 Tiemarifing LCV 2.5 Sumalemba MV 2.7

 

The frequencies in this table don’t add to 100 because there were more than 10 varieties (28 in Dioila and
29 in Maude) listed by farmers as dominant preferred varieties. The percentages were taking from the total
sample of sorghum growers interviewed (79 in Dioila and 75 in Maude)

LCV refers to local varieties used as control varieties in trials, FV 07/08 refers to farmers varieties included
in trial for 2007/2008 and MV stands for modern varieties

The results in table 3-8 indicate that in both zones, varieties such as Seguetana,
Tieble were cited by farmers as their preferred dominant varieties i.e. the varieties they
plant in larger area because they like them.

Table 3-9 presents the results regarding farmers’ preferred varietal trait evaluation

from the farmer survey conducted in both zones. Results from women were not reported

113

in this evaluation because it was conducted with sorghum growers in the survey sample,

of which women represented only 5%.

Table 3-9: Results of farmers’ variety traits preference evaluation of their preferred
dominant varieties expressed in percentage, from the farmers’ general survey (the initial
survey) conducted in December 2007 in the Dioila and Maude zones in southern Mali

 

 

Varietal traits Percentage of farmers who listed the trait
Yield 95
Resistance to insect 58
Early maturity 43
Grain cooking traits 75
Drought tolerance 72
Secondary uses (like animal feed) 17
Other 15

 

The percentages summed are greater than 100% because farmers were allowed to cite as many variety traits
as they thought were important.

Table 3-9 shows that yield, cited by 95% of the farmers, is the most preferred
varietal trait by farmers in both zones. Cooking characteristics, listed by 75% of farmers
interviewed, occupies the second place. Finally, drought tolerance, cited by 72% of the

farmers, is the third most important varietal trait in the areas.

Results from farmers’ survey per zone

Table 3-10: Results of farmers’ variety traits preference evaluation of their preferred
dominant varieties, expressed in percentage per zone from the farmers’ general survey
(the initial survey) conducted in December 2007 in Dioila and Maude zones.

 

 

Varietal traits Dioila n=79 Maude n=75
Yield 97 92
Resistance to insect 67 49
Early maturing 58 28
Grain cooking traits 91 57
Drought tolerance 76 68
Secondary uses 25 8
Others 10 20

 

The percentages summed are greater than 100% because famiers were allowed to cite a variety traits as
many time as they think important.

”4

From table 3-10, the three most popular varietal traits preferred by farmers in both
zones are the same. Yield is the number one in both zones, with 97% in Dioila and 91%
in Maude. There is slight switch in position for cooking characteristics and drought
tolerance. While cooking characteristics (91%) occupied the second place and drought

tolerance, with 76%, the third place in Dioila, in Maude their position switched.

115

 

DISCUSSION

Farmers have always been plant breeders, although they are often not formerly
recognized as such at the institutional level (Chakanda, 2000; Almekinders,1999;
ICRISAT, 2000). Farmers’ choice of varieties is based on their farming objectives, the
environment and farming conditions where they live (Diakite, 2003; Kudadjie et al.,
2004). Under irregular and low rainfall farming conditions, early maturity or drought
tolerance, in addition to yield, are expected to be preferred variety traits; however, the
evaluation by farmers in this study did not provide much evidence for interest in early
sorghum varieties.

Overall, farmers participating in PVS field trial evaluation and in the survey were
interested in similar varieties, and specifically in traits such as high yield potential,
adaptation to the field enviromnent and grain appearance (Tables 3-5 and 3-6). These
results indicate that sorghum breeders can breed for varieties that address farmers’
preferences across this region, without having to tailor a specific variety for each zone.
The farmer interest in grain appearance documented here is in accord with the PVS
literature that shows farmers have strong preferences for specific grain color and size
aspects (Diakité 2003). This is somewhat surprising in that early maturity and drought
tolerance would appear to be of increasing importance, as climatic variability in Mali
over the last decade has included drought, and short and irregular rainfall (Simpson
1999). An earlier study conducted in the region of my study by Siart, (2008) found
evidence that early flowering was related to early maturity and high yield, which were the
variety characteristics with the highest frequency in farmers’ ranking. These traits were

followed by culinary qualities, which is a similar characteristic to the cooking traits

116

described by farmers in our study (Table 3-8). In contrast to that the research reported
here, visual traits such as grain appearance were mentioned by only a few farmers (<7%)
in Siart’s research.

The largest difference between the results we observed from farmer evaluation in
PVS trials and the survey results were that grain appearance was important in the PVS
variety evaluation (table 3-4),Iwhile size and color traits were not mentioned by farmers
who we surveyed. This might be related to the setting of the evaluation. When evaluating
16 varieties in a field trial, clear differences can be observed in terrus of grain color and
size in addition to yield. Since the fields were all at maturity, the grain appearance was an
obvious factor. This result is consistent with Baidu-Forson’s (1997) study on “on-station
farmer participatory varietal evaluation” of millet in Niger. The setting could help explain
the result by gender as well, because the presence of men could influence women’s
evaluation and criteria, and could have focused attention on yield rather than post-harvest
characteristics. Women’s preference criteria were supposed to be different from men’s
including processing and cooking traits, but this was not observed here (Table 3-6).

The results of the farmers’ survey showed that yield, cooking characteristics, and
drought tolerance were reasons that farmers preferred specific sorghum varieties. These
results accorded with our research hypothesis, which stated that the three most dominant
varietal traits would be yield, early maturity and drought tolerance. It is also close to
Siart’s (2008) finding on farmers’ varieties choice in the two zones. These results agree
with the results of Chakanda (2000) and the results of Nkongolo et al.. (2008) on PVS
and characterization of sorghum in Malawi, indicating that farmers prefer varieties

adapted to their climatic conditions, production goals and specific utilization and post-

117

harvest characteristics (Siart, 2008). Yapi et al. (2000) showed that farmers’ choice for
new sorghum varieties was determined by their socio-economic, environmental and
climatic conditions; where farmers’ preferences in all their study regions (Segou, Mopti
and Koulikoro regions) for sorghum and millet were for early maturity, yield and food
quality. .

Since sorghum is mainly produced in Mali as a staple food crop, it is not
surprising that the survey elicited farmer criteria that included processing and food
quality as a determinant factor in acceptability by farmers, as indicated in Diakite (2004)
in his study on improved sorghum diffusion in Dioila and Maude zone. Kudadjie (2004),
in his diagnostic study on assessing production constraints, management and use of
sorghum diversity in north-east Ghana, also found similar result that farmers’ preferences
for different sorghum varieties depends on agronomic and gastronomic variety traits. A
PVS with improved pearl millet study done by Omanya et al. (2006) in West Afiica
(Mali, Niger, and Burkina-Faso) showed that early maturity was the second important
variety trait for millet farmers followed by adaptation and acceptable taste. Their results
indicated that farmers’ preferences for crop varieties to plant are influenced by their
farming goals and environmental conditions, especially the quantity of rainfall.

A deeper look of the data showed that farmers in Dioila rated cooking
characteristics higher than drought tolerance, while Maude farmers switched the ranking
by placing drought tolerance before cooking traits. We conducted a follow-up workshop
in each village, where initial results were discussed with farmers, and documented
through this discussion that although a variety called “SUMBA” is high yielding and

early maturing, some farmers in Dioila said they did not prefer it because of its poor

118

cooking quality (author’s follow-up workshop in Dioila zone, 2008). A farmer from
Wakoro said: “S UMBA is a very high-yielding variety, and it matures earlier than all our
sorghum varieties, but I will never plant it again in my field because my wives complain
about its processing and the quality of the-food. I sold all my production to a grain seller
I knew. The following year, he told me to not bring again this type of grain because he
almost all his clients complained about it " (September 12th 2008 workshop in Dioila).
Some farmers in the same zones, however, said that they like SUMBA because its stems
and grains are very good animal feed. A farmer from Magnambougou said “ My wives
don 't have problems processing and cooking S UMBA because there are some dishes like
“Yeyekiny and Bashy " (cous-cous) that can be made well out of it " (September l3"I

2008 workshop in Dioila).

119

CONCLUSIONS

A finding from both the on-farm field trial evaluation and the survey was that
yield and adaptation to the local environment were very important variety evaluation
criteria, but there were additional traits of interest to farmers from Dioila and Maude.
These included cooking traits (easy processing, good taste and easy storage of dishes
made of a variety) and drought tolerance. However, the farmers in Dioila were more
interested in cooking quality than in drought tolerance. In some cases, they even put
cooking quality before yield. For example, the variety SUMBA was not preferred by
some farmers, despite it high yield and early maturing, as it had poor cooking quality
traits for the locally preferred ‘to’.

A two-component evaluation of farmers’ decision criteria and preferred varietal
characteristics provided additional insights. The PVS trials provided opportunities for
farruers to examine and comment on specific varieties in a group setting, whereas the
individual surveys provided more opportunity for comments on post-harvest traits, and
reduced the influence of farmers on each other’s criteria. The results of these evaluations
will help researchers to develop varieties that meet the majority of farmers’ needs in a

given area.

120

CHAPTER 4

IMPROVING FARMERS SEED PROVISION FOR
SORGHUM

INTRODUCTION

Seed is a central part of a farmer’s life (Leonardo, 2002). Seed management is a
central issue for farmers and a key element in addressing the challenges of responding to
farmers’ different requirements and preferences, increasing agricultural production, and

achieving food security in Sub-Saharan Africa (SSA) (Almekinders et al. 1994).

Understanding seed demand and supply among small-scale farmers is particularly
crucial for crops that are not high value, such as subsistence pulses or cereals, which are
also are self—pollinating crops (e. g. the common bean, groundnuts, sorghum...) (David,
2003). In order to improve farmers’ access to good quality seeds, it is imperative to
understand the existing seed system to be able to identify the weaknesses and strengths as
a basis for developing more well functioning seed systems (Siart, 2008).

There are several factors that influence farmers’ seed provision—for exaruple the
need to replace the seed of an existing variety that is no longer meeting their farming
goal, or to obtain a new variety with higher yield or more pest and/or diseases resistance
(Heisey, 1991;Tripp, 1997). Seed provision can then be divided into two categories.
Seeds from the formal sector, commercial or certified seeds, are supplied by commercial
input distributors and produced under strict conditions to meet the certification
requirement. Seeds from the informal sector include farmer-saved seeds, farmer-to-

farmer exchange seed, and purchase of market grain for use as seed (Pejuan, 2005). In

121

most developing countries such as Mali, smallholder farmers rely primarily on their own
production (saved-seed) as the source of seed for subsistence crops such as sorghum
(Siart 2008). Many farmers in Mali have few resources, with limited access to
agricultural inputs, services, and markets (Setimela et al., 2004).

In 2006/07, farmers in both research zones of this study (Dioila and Maude)
obtained sorghum seeds from three sources: their own stock, saved from the previous
season, local markets and other farmers (author’s survey data). This pattern is common
for most subsistence crops including millet, groundnut in the research areas.

Worldwide, local seed systems are poorly understood, and few empirical studies
on this topic exist (Cromwell, 1990, Almekinders et al., 1994, David, 2003, Siart, 2008,
Christinck, 2002). Local seeds are important in relation to on-farm crop diversity and for
meeting the local seed requirements of farmers (Leonardo, 2002).

Literature on seed systems and technology adoption in developing countries offers
various explanations for non-adoption of improved crop varieties by farmers (Tripp,
1997, David, 2003). Two of the more common explanations that are relevant to improved
varieties adoption include (a) lack of information about improved varieties and/or lack of
access to improved seed and (b) inappropriate or unprofitable technology (i.e.,
experimental results are not representative of farmer agronomic and/or economic
conditions) (Pejuan, 2005). Other explanations for weak technology adoption for
sorghum in Mali are the cost attached to the adoption of the new varieties, the subsistence
nature of sorghum production and the availability of sorghum seed in informal sector

(Yapi et a1, 2000; Siart 2008; Almekinder and Lourwaars 1999). The high cost of seed

122

associated with formal seed systems is in part due to regulations and a complicated
variety release process in many SSA countries (Maredia et al., 1999).

Seed system analysis in Mali has shown that farmers frequently change varieties
they sow and are interested in trying new types of varieties that include their preferred
variety traits (ICRISAT, 2005; Siart 2008). Today in Mali, there are regulations on the
books that could complicate release of new varieties and consequently access to seeds for
farmers, but in practice they are not enforced.

Insights into seed system functioning may be obtained by assessing farmers’
willingness to pay for seeds that meet their needs. Understanding seed systems, and the
role of seed fairs, could contribute significantly to the development of good quality seed

production and marketing systems in Mali.

Sorghum seed marketing in Mali

ICRISAT has supported farmer organizations to form farmer committees at
community level. In the Dioila zone, the farmer committees were from OPC'zs of ULPC,
and in the Maude zone, they were formed with help of ACOD and AOPP into a
cooperative called COPROSEM (Weltzien et al., 2007). In both zones, some members of
these farmer committees hosted on-farm field trials with ICRISAT and IER, and then
started producing seeds locally after training by the research technicians (Weltzien et al.,
2006). Local seed producers through the farmer committees organize seed sales in their
community, as well as working through seed fairs in the Maude zone and OPCs stores in

Dioila.

 

'2 OPC: Organization of cereal producers

123

Previous research has indicated that farmers’ demand for seed of modern varieties
of sorghum (MV) depends on the performance of varieties in farmer’ fields compared to
their traditional varieties (Pejuan, 2005), including their adaptation to local soil and
climatic conditions. In addition, demand for purchasing seeds will depend on the
availability and affordability of seed.

Sorghum seed sales were a relatively new concept to farmers in this region of
Mali. In the Maude zone, seed fairs were initiated in 2005 and in Dioila with the OPCs in
2003-2004 (Siart, 2008). This can be an explanation for the limited sourcing of seed from
markets in both zones. There are socio-cultural reasons as well, as farmers do not
consider it appropriate to sell or buy sorghum seed (Siart, 2008). Seed distribution
activities must take into account that seed selling from farmer to farmer is not socially
acceptable. Broek (2007), in her survey of women in these two zones, also found weak
market demand for sorghum seed, which she attributed as being due to poverty, lack of
information on the new varieties on the markets and the existence of seeds through
informal systems such as trading seed through kin and neighbor networks.

As most farmers have limited access to transportation (Setimela et al., 2004),

these constraints limit access to purchased seeds (David, 2003).

Seed fairs

Seed fairs were defined by Almekinders et al., 1999 as meetings where farmers
trade seeds, exchange information about their varieties and crop species, and share their
knowledge about agro-biodiversity management. In Zimbabwe, community seed fairs
were used as an approach to facilitate access to and use of diverse and locally produced

seeds to promote local seed security (CTDT, 2006). The main achievement of the seed

124

fairs was increased crop genetic diversity at the community level and greater capacity
among farmers to judge and select plants and thus to make informed decisions in
breeding (CTDT, 2006).

Seed fairs were first introduced in the study area with farruer seed producer
cooperatives around 2005 as an application of one of the recommendations from the seed
system security assessment done in northern Mali by Sperling et al. (2006). The first test
of seed fairs in Mande zone was held in May 2005 and was judged as successful by the
organizers based on the participants’ appreciation of the seed quality, price, the setting

(place and day) of seed fairs (Siart, 2008).

Willingness to pay

Myrick (1993, Pejuan, 2005) defines willingness to pay as the maximum amount
of money an individual is willing to pay instead of doing without increase in the quantity
of some good. Willingness to pay is most frequently associated with nonmarket valuation
techniques like contingent valuation (Hanley et al., 1998), it can also be used with
observed data of marketed good. Farmers’ willingness to pay an agricultural input such as
seed of modern sorghum varieties depends on its price, their economic situation, and the
availability of credit and/or government intervention such as subsidies, but also on the

additional gain or advantages expected from using the seed.

Objectives

The main objective of this chapter is to investigate ways to improve farmers’
access to good quality seeds on time and at affordable prices of their preferred varieties.
Specific objectives:

0 Assess farmers’ willingness to pay for seed of preferred sorghum varieties

125

0 Assess the suitability of seed fairs as a way to improve seed distribution

and to promote the dissemination of improved sorghum varieties

Hypotheses

H1: Farmers are willing to pay for seeds of preferred sorghum varieties

H2: The price farmers are willing to pay depends on variety traits

H3: The price farmers are willing to pay is high enough to attract seed
companies or COOPROSEM or other individual seed producers to invest in the
sorghum seed business.

H4: Seed fairs are useful tools to promote sorghum seed of improved

varieties

126

MATERIAL AND METHODS

To assess the usefiiluess of seed marketing as way to improve farmers’ access to
good quality seed, I used the willingness to pay approach and an evaluation of seed fairs
done in the Maude zone. Both the willingness to pay and the seed fair evaluation were

done through surveys.

Seed fair survey

In May of 2007 we conducted surveys with sellers, buyers, and visitors at two
seed fairs organized by COOPROSEM (local union of farmer seed producers) in the
Maude zone, in Bancoumana on 5/14/08 and in Siby on 5/12/08. The goal of our survey
was to see how well seed fairs function in terms of improving farmers’ access to modem
varieties in rural areas of Mali. In the first part of the survey, seed sellers were asked
about the species and varieties (local or modern) of seeds they brought to the seed fairs to
sell. For each crop (or variety), we asked about the source of seed and whether it was
self-produced or from others (friends or relatives who could not attend the fair). We were
interested also in determining the sellers’ experience in seed selling as well as their
participation at previous seed fairs. This was to understand their motivations and
coucems regarding participating in seed fairs. We were interested in detenniuiug if seed
fairs were effective means of selling seed, from the perspectives of seed sellers.

The second part of the survey we interviewed seed buyers to investigate how
effectively the infonnatiou about seed fairs was promoted in rural areas. In addition to
this information, we wanted to know why these farmers decided to buy seeds at seed
fairs. We also asked the buyers their impressions regarding the seed quality, prices,

presentation (size of bag/label), and varietal diversity available at seed fairs. Finally since

127

seed sellers (COOPROSEM, members) are trained by ACOD and AOPP agents in
marketing seeds, I wanted to know whether the information that they (seed sellers and
NGOs workers) provided was helpful in terms of buyers’ choice of varieties. If the
inforrnatiou was not helpful or was insufficient, we asked what further information
buyers would like to have, and what would they be interested in seeing improved about
the seed fair to meet their needs.

The survey conducted of fair visitorsl3 ’ was to ascertain the following: a) Why
did he/she decide to visit the seed fair? b) How had they heard about it? c) Did they ever
buy seeds in any seed fairs? If yes, what (species), what varieties, how many times and
where they did they buy?. Then why he/she did or did not buy anything at the fair today?
An open-ended question was asked of visitors regarding their overall impression about

seed fairs in terms of place (location), time (day in the year), and a general rating.

Willingness to pay survey

This part of the data collection was done as part of the evaluation of farmers’
preferred sorghum varieties (reported on in chapter 3). The survey is described in detail
earlier. In brief, it was conducted in 9 villages (five in Dioila and four in Maude) and
focused on farmers’ rationales for their choice of the three best varieties. At the end of
the survey, they were asked about how much they were willing to pay for the seeds of
their top three preferred varieties if the market was the only seed source for this seed.. A
baseline price was given to the farmers which they could not go below; this was the price

of grain for consumption. The grain price, used as the base price in my study, was 150

 

'3 Visitors are people who came to the fair just to look around and didn’t buy anything.

128

FCFAM/kg. For example, for the number one choice variety, a farruer could be willing to
pay 200 FCFA/kg, second best = 155 FCFA/kg and 150 FCFA/kg for the third.

This survey was conducted in collaboration with three ICRISAT technicians
(Arouna Sangare, Moussa Kanoute, and Sidy Dembele) and two IER technicians (Bocar
Diallo and Karim Cisse). These technicians were experienced at evaluation and
conducting surveys on farmer preferences, as they constituted ICRISAT/IER on-farm
farmers’ field trial evaluation team, where they provided technical assistance in

participatory plant breeding trials hosted by farmers in this region.

General farmers’ survey

This survey is part of the general farmers’ survey on seed quality assessment
described in chapter 1. It provided an opportunity to document farmers’ seed provision
(source, mode of acquisition, time of acquisition and reasons of choice of the source). For
all information about the process, the sample size, the time and places (villages) the
survey was conducted, please refer to chapter 1. Also, the survey materiel is provided in

the appendix of the document.

Data documentation and analyses

The data for this chapter were collected by me and three people, including two
local NGOs workers and one extension worker of a governmental agricultural service
(Mamadou Coulibaly from the “Association des Organisations Paysannes des
Producteurs” (AOPP), Adoulaye Sangare from the“Association Conseil pour le
Développement” (ACOD) and Brehima C amara from the “Office de la Haute Valle’e du

Niger” (OHVN)).

 

'4 FCFA is the currency used in Mali, at the survey time IUS dollars was 450 FCFA

129

The data were first recorded in notebooks and questionnaires in the field. They
were then entered in a laptop computer in Microsoft Office Excel. The data were
analyzed using Microsoft Office Excel as well for standard descriptive statistics (means,
maximum , minimum prices, and standard deviation of the sample) and frequency

analyses calculation.

130

RESULTS

Results from the general survey
Seed sourcing

Table 4-1: The main sources of sorghum seed, villages where seeds were obtained, and
the time of its acquisition by zone (Dioila and Maude), from the farmer general survey
done on December 2007

 

Dioila n=121 Maude n=120
Seed sources in percentage in the sample per zone

 

Own production 71 85
Market 10 7

Relatives l6 8

Research 2 0

Others 1 0

From Where (%) (the village where the seeds were obtained)

Same village 85 84
Another village 15 16
Time of acquisition (”/0 in the samples per zone)

Since harvest 49 87
Planting time 42 9

Others 9 4

 

Farmer’s own production - saved from a previous harvest - was the main source
of sorghum seed in both zones, 71% in Dioila and 85% in Maude (table 4-1). The market
was the third source in both zones, 10% in Dioila and 7% in Mali.

Looking at the villages where farmers got their seeds, the results from these data
showed that in both zones, most seeds came from the same village where the farmer lived
(85% for Dioila and 84% for Maude).

Most seeds in both zones were obtained since harvest. In other words, during
harvest or just after harvest, farmers set apart the quantity of sorghum that they intended

to use as seed for the next planting season. However, more farmers in Maude (87%)

131

chose their seeds at harvest time than Dioila (49%). In the Dioila zone, 42% of the

sorghum producers interviewed got their seeds at planting time.

Market sourced sorghum seeds from the general survey

Table 4-2: Information on sorghum purchased seeds in Dioila and Maude, from farmer
general survey conducted on December 2007 to document market sourced seed in both
zones.

 

Dioila n=14 Maude n=8

 

Place of purchase

Market 0 100
COOPROSEM/ULPC ' 100 0
Choice of seller

Familial relationship 0 50
Information from other growers 100 50
Average Distance to seller in km 10 20
Type of seller used

Seed producer association 100 0
Grain sellers 0 100
Mode of payment

Cash 100 100
Reason of choice for a purchased source

Lack of seed 50 50
Past experience with the varieties 50 50
Satisfaction with this source

Yes 100 100
No

 

The results from table 4.2 show that farmers’ seed producer association stores
were the place where all purchased seeds (100%) in Dioila were from, while in Maude,
the market place (terminal market) was the place where 100% of purchased seeds were
obtained.

In Dioila, the choice of seller (from whom to buy) was 100% based on
information obtained from other sorghum growers. In Maude, on the other hand, the
choice of seller was 50% based on familial relationships and 50% on information

obtained from other sorghum growers in the zone. The type of sellers used by seeds

132

buyers in Dioila was 100% seed producer association and 100% grain sellers in Maude.
Buyers in Dioila traveled in average 10 km to get their seeds, while buyers in Maude had
to travel 20 km on average to get their seeds. Cash payment was the mode of payment
used in both zones.

Lack of seeds and the farmer’s past experience with the varieties were the two
main reasons for choosing market seed source in both zones. All the buyers reported that

they were satisfied with the seeds they purchased.

Result from farmers willingness to pay survey
Farmers willingness to pay for sorghum seeds in Dioila and Maude zone

Table 4-3: Prices (in F CFA) that farmers are willingness to pay for seeds of their three
most preferred varieties, for the entire sample (n=230).

 

 

Prices in FCFA/kg lrst choice 2cd choice 3rd choice Average
MEAN 245 193 169 202
MAX 2000 1000 500 1 l 67
MIN * 75 50 30 52
STDV 175 87 60 107

 

"‘ The minimum price reported in this table is below the 150 FCFA reported in the method section because
no matter how hard we insisted to not propose price below that some people would give less than 150

As presented in table 4-3, farmers are willing to pay on average 202 F CF A/kg for
seed of their top three choices sorghum varieties. However, with an average maximum
price of l 167FCFA/kg, an average minimum of 52 FCF A /kg and standard deviation
around the mean of 107 FCFA/kg, the average price was subject to a high variability.
The location of the farmers influenced their willingness to pay, and thus was a
contributor to this variation in price. As shown in Table 4—4, farmers from Maude were

on average willing to pay 38% more than were farmers from Dioila.

133

Table 4-4: Farmers’ willingness to pay for their preferred varieties survey per zone

 

 

Prices in
FCFA/kg Dioila n=141 Maude n= 89
Average price 176 244
Max 300 1167
Min* 52 83
STDEV 40 155

 

* The minimum price reported in this table is below the 150 FCF A reported in the method section because
no matter how hard we insisted to not propose price below that some people would give less than 150

Further, the standard deviation in Maude was very high compared to that in
Dioila. This result suggested that the average price of Dioila provides a better inference

about the preferences of all the sorghum growers in Dioila than does the average price of

Maude.
One hundred twenty four men and 106 women participated in the variety
evaluation exercise. There was no influence of gender on willingness to pay; this is

shown in figure 4-1 where the variability was high and there was no significant difference

by gender;

500
450 a
400 e
350 e
300 *
250 e
200 1
150 .
100 4
50 i

 

 

1:] Men
I_Women

 

pricedakilod‘seedin FCFA

 

Dioila Mande
Zones

 

   

 

 

l _..- ‘___, L_ v _ fig _ k

Figure4—l: Farmers willingness to pay evaluated by gender in Dioila and Maude zones
through the survey done along with on-farm variety evaluation in the two zones in
December 2007

134

Farmers ’ willingness to pay for variety traits

Table 4-5: Results of farmers’ mean willingness to pay for their preferred sorghum
varieties in relationship to the most popular variety traits studied in chapter 3. Mean price
in CFA for a kg of seed in relationship to varietal traits (the standard deviations are in
parentheses)

 

lrst choice 2cd choice 3rd choice Average

 

Yield 241 (109) 193 (74) 169 (57) 201 (80)
Early maturing 220 (63) 181 (37) 185 (58) 195 (53)
Drought and Striga tolerance 239 (105) 186 (40) 161 (32) 195 (59)
Grain appearance 244 (125) 193 (72) 168 (46) 202 (81)
Adaptation to the farming areas 225 (52) 176 (28) 161 (41) 187 (40)
Others 207 (70) 176 (34) 160 (74) 181 (59)

 

The results in the table 4-5 show that farmers are generally willing to pay more
for their top (number one) preferred variety, but this varies in relation to variety traits,
and the prices for the top variety are subject higher variability than the two other
preferred varieties. For the first preferred variety, the ranking of the willingness to pay
puts grain appearance in first place, with an average price of 244 F CFA/kg; followed by
yield, with a mean price of 241 FCFA/kg; and resistance to drought and Striga in the
third position, with a mean price of 239 F CFA/kg. For the second preferred variety, grain
appearance and yield tie for first rank, with an average price of l93FCFA/kg; followed
by resistance to drought and Striga, with 186FCFA/kg. For the third choice, earliness got
the highest price (1 85FCFA/kg) ,followed by yield (l69FCFA/kg) and grain appearance
(168 FCFA/kg).

For the overall evaluation, grain appearance and yield got the highest willingness

to pay, but the proposed prices have higher variability than the other traits.

135

Evaluation of the seed fairs

Table 4-6: The numbers of participants interviewed at the fairs held in Bancoumana and
Siby in May 2007 in Maude zone to test the appropriateness of seed fairs as tool to
improve farmers’ access to good quality seed of their preferred varieties.

 

 

Sellers Buyers Visitors total
Siby 1 0 23 3 l 64
Bancoumana 12 34 3 l 77
Total 1 2* 57 62 1 41

 

"‘ The total of sellers being 12 instead of 22 is because it was the same sellers from Siby plus two grain
vendors who were in Bancoumana

Among the interviewed participants, about 47% of them were visitors and 43%
were buyers. The 34 buyers in Bancoumana could be due to the fact that the fair of Siby
happened first, and provided background information that helped farmers choose to buy
their seeds at Bancoumana.

Table 4-7: Numbers of crops species presented at the seed fairs held in May 2007 in
Bancoumana and Siby in the Maude zone.

 

Sorghum* Groundnut Rice maize Cowpea Millets
Siby 6 2 6 0 l 0
Bancoumana 8 0 0 2 0 0

"' The modern sorghum varieties at the fairs were ICRISAT or IER varieties in production at the farmers’
level. They are produced by members of the cooperative of farmers producing seeds in that Maude zone
that get technical assistance from ICRISAT, ACOD and AOPP (COOPROSEM) to sell to the other

farmers.

 

 

136

Table 4-8: Information on crop types (number of varieties, variety names, variety types
and reasons of selling it) presented on the seed fairs held in Bancounama and Siby in
May 2007

 

Number Variety
Crop types varieties Variety name types Reason for selling this variety

 

Sakoykaba Modern High yield
Soumalemba Modern High yield and farmers like it

Toroba Modern First time of being sold
Weli Modern Not being sold yet
Girinkan Modern Not being sold yet
Sorghum 9 Niangaye Modern Not being sold yet
Tieblen Modern High yield and good taste
Toronkanikela local Not being sold yet
High yield and farmers like it
Kalaban Modern
Nerical Modern High yield
Nerica2 Modern High yield
Nerica 4 Modern New variety in diffiJsion
Rice 6 Sikassoka Modern New varieties in diffusion
Kumbabaui Modern New varieties in diffusion
J igifa Modern New varieties in diffusion
High yield and good source of
Cowpea l Cowpea local cash
Birindima Modern High yield and early maturing
High yield and adapted to the
Groundnut 2 Ti gaba local area
High yield and adapted to the
Maize 1 Sotubaka Modern area

 

Sorghum and rice were the dominant crops with seed offered for sale at the fairs.
Among the nine varieties of sorghum, 8 were modern varieties. These varieties are
ICRISAT or IER varieties developed through PPB, and their seeds are produced by the
members of COOPROSEM, who with help of ICRISAT, ACOD and AOPP organized
the fairs. These sorghum varieties were chosen for seed production, because farmers had

shown interest in them during the ou-fann variety evaluations.

137

Table 4-9: Information collected from seed fair vendors about their experience as seed
sellers and membership in farmers’ organizations

 

 

Information number in the sample percentage
Member of COPROSEM l l 92
Grain vendors I 8
Participation in previousfairs

Yes 1 1 92
No l 8
Participation in previous fairs as vendors

Yes 10 92
No l 8
Motivationfor participating in the seed fairs ("0)

Diffusion of new varieties 17
Discover new seed production techniques 25
Become member of COOPROSEM 8
Get financial and technical assistance from research and extension

services 8
Bring more seeds in the fairs 25
Others 17
Satisfaction about the seedfair (organization andfunctioning)

Yes 11 92
No 1 8

 

As shown in table 4-9, the sellers at the Maude seed fairs were predominantly
(92%) members of COPROSEM. Among the COPROSEM seed sellers, all had
participated in previous (2005 and 2006) seed fairs held in the Maude zone, and all but
one were seed sellers previously. Some of the sellers (25%) were motivated to participate
in the fairs to discover new seed production techniques and to bring more seeds to the
fairs. Others were motivated to participate in the fairs to help diffuse new sorghum
varieties (17%), or to become member of COPROSEM (8%) and to obtain financial
and/or technical assistance from research and extension services (8%). Almost all the

sellers (92%) indicated that they were satisfied with their participation in the fairs.

138

Table 4-10: Information collected from seed fairs seed buyers about their experience in
buying seed and the reason why they chose to buy seed at the fairs

 

 

Percentage
Gender
Men 77
Women 23
Types of farmers
Normal farmers 88
Tester of new sorghum varieties 0
No response 12
Information about thefairs
Market 35
Radio 31
Other farmers 23
Others 1 1
Reason for buying seeds
Interested in new varieties 19
Variety characteristics 66
Test 4
Seed quality 4
No response 7

 

The results presented in table 4-10 show that 77% of the seed buyers were men
and 23% women. The majority of farmers (88%) were not involved in plant breeding
research, e.g., did not host on-farm field trial programs. About one-third (35%) of the
buyers had heard about the fairs from the market”, 31% learned about the seed fair from
a local radio program, and 21% were informed by other farmers. Among the buyers,
66% cited variety characteristics as their reason for buying seeds in the fairs; some 19%

bought because they were interested in testing new varieties.

 

'5 When they came in the market, they heard the sound of the drum and went to see what was happening.

139

Table 4-11: Information collected from seed fairs visitors about their experience in
visiting the fairs held in Siby and Bancoumana on May 2007 in Maude Zone

 

 

Percentage
Connection with COOPROSEM or [CR/SA T
Normal farmer 89
Tester 1 1
Motivation for visiting the fairs
Just look 13
Get information on new varieties 77
Others 8
Source of information about thefairs
Market 18
Radio
Project workers 6
Other farmers &radio 8
Radio & project workers 5
Radio-proj ect workers-other farmers 6
Others 28
Have you ever bought seeds in afair?
Yes 8
No 92
Why you didn 't buy anything to day?
No money 43
Already have seeds 6
Will buy another day 37
Others 14

 

From table 4-1 1, 89% of the visitors were farmers who did not participate in
research trials, and 77% of the visitors stopped at the fairs to get information on new
varieties that they had seen with the seed producers. Most of these visitors heard about
the fairs from local radio programs (29%), other sources (28%) and from the market
(18%). The other 25% heard from project (ACOD and AOPP) workers, or other farmers,
or radio and project workers, and/or radio-project worker. Ninety-two percent of the
visitors had never bought seeds at a seed fair, but they took the time to stop at the fair.

Among the visitors, 43% did not buy seed because they did not have money that day, and

140

37% said they would buy seeds another day. (They needed time to go look for money or
make a decision about what varieties to plant.)

Table 4-12: Quantity of sorghum seed purchased at the seed fairs held in Bancoumana
and Siby in May 2007 during the time the surveys were conducted

 

 

Variety names Quantity in kg
Bobodje 4
Kalaban 2
Sakoykaba 4
Sumalemba 1 6.5
Tieblen 3

Total 27.5

 

The numbers presented in this table do not reflect the total amount of seeds
purchased during the fairs because we did not stay until the end of the fairs.

Table 4-12 shows that Sumalemba was the most purchased variety during the
fairs. Alone, it accounted for 60% of the total quantity of seeds purchased during the

period of the survey. The quantity purchased per buyer ranged from 0.5 to 2 kg.

141

Participants impression on seeds at the fairs

Table 4-13: Buyers’ impressions on seed quality, prices and seed presentation in the seed
fairs

 

 

Percentage
Seed quality
Very good 12
Good 77
Poor 0
Don't know 1 1
Seed prices
Very high 4
High 35
Acceptable 42
Low 4
Don’t know 15
Seed presentation i. e. packaging
Good 73
Bad 7
No answer 20
Obtaining the preferred varieties
Yes 88
No 12

 

The results presented here show that seed quality was judged good for 77% of the
buyers and very good by 12%. Forty-two percent of the buyers also found that seed prices
were acceptable, while 35% found that they were high. Fifteen percent of the buyers
didn’t want to answer this part of the questionnaire. In terms of seed presentation, i.e.,
packaging, 73% of the buyers said it was good, and 20% didn’t want to comment on that
part. In terms of ability to obtain their preferred varieties, 88% said they were able to do

SO.

I42

Participants ’ suggestions for future improvement of the seed fairs

Table 4-14: Suggestions made by the interviewed seed fair participant in Bancoumana
and Siby for future improvement of seed fairs in the Maude zone.

 

Seed sellers

Buyers

Visitors

 

Get fixed places where
farmers can get seeds after
the fairs

Have more crops and
varietal diversity so that
everybody can get what
they want

Help the COOPROSEM to
widely inform farmers on
the variety traits of the new
varieties and the seed
quality

Decentralize the places of
the fairs, i.e., hold the fair
in more places

Bring more crops
and varieties
Reduce the price of
seeds

Spread the
information about
the fairs frequently
and on time
Provide more
information on the
varieties

Have the fair every
year at the
beginning of rainy
season

Make some
mobilization
campaign on seed
quality

Have seed fairs more
than once a year
Add more crop types
and varieties

Reduce the price of
seeds

Associate all the
local farmers
organizations

Give more
information about
the fairs, the crops
and varieties

Make more
advertisements on
local radios
Diversify the place
of the fairs

Choose a place
easily accessible like
the current places

 

All seed fair participants interviewed shared similar coucems in terms of interest

in having access to more crop types and varieties at the fair. Another common point was
to have more information about the fair, and more frequent fairs held at more locations.
For promotion of the fair, they suggested that it be announced in local radio programs and
broadly disseminated using different channels. There was also interested in information
being made available about varietal characteristics associated with the varieties the sellers

bring to the fair.

143

DISCUSSION

Major elements were highlighted by this study regarding sorghum seed systems
function in Mali. These included a high level of farmer-produced seed (80%), local
sourcing of seed, and limited adoption of improved varieties (Tables 4-1).As found by an
earlier survey conducted in the same zones (Diakite, 2004), more than 80% of the seeds
were from the village in which the grower lives.

New sorghum varieties have been released with higher yield potential and good
cooking habits especially Tieble , Ngologing, Fambe, Soumalemba, Soumba and
Sakoykaba In both of the study zones, these new varieties were favored because of their
yield, relative earliness, excellent grain qualities and adaptation to 800-1000 mm rainfall
zone (Siart, 2008). Despite farmer preference for these traits associated with some of the
modern sorghum varieties, there still was almost no market for sorghum seeds by which
farmers could access these varieties in the study zones. Own seed production is still the
dominant seed source (>70% of the total seeds used) for sorghum in both zones (Table 4-
1), which is similar to earlier findings from surveys conducted in Dioila and Maude by
Diakite, 2004, Siart et al., 2005, and Siart et al., 2006.

Forty-two percent of sorghum seeds in Dioila zone were obtained at sowing time.
This result can be related to farmers’ seeds management practices in Dioila. Siart, 2008
showed that 71% of sorghum in Maude was stored in sheavesI6 selected during harvest in
the field; this percentage was only 30% in Dioila. This means that many farmers in Dioila

don’t separate seeds from grain. It is only at sowing time that they start looking for seeds

 

'6 A sheave is a group of panicles selected and attached together by farmers at harvest time to be used as
seeds for next planting seasons.

r44

from their grain, and in case all their production was consumed, they will then use other
sources (market, friends and relatives). Our result also showed that more farmers in
Dioila purchased seed, and these farmers in general prefer to buy just at sowing time,
which confirms earlier findings from this area (Siart, 2008). Diakite (2004) found that
there are part-time farmers (e.g., school teachers or NGO staff) in Dioila zone who do not
keep their own seed regularly and are keen to buy good quality seed on a regular basis.
Our results for market sourced seeds show that in Dioila seeds were purchased
from the OPCs, while in Maude, seeds were purchased at markets. There were two seed
fairs held in Maude, so that it is surprising that the survey found no purchased seeds.
However, Diakité 2004 showed that most farmers in the study areas obtained new
varieties from friends. Thus in Maude, most of the farmers who want the new varieties of
COOPROSEM might have gotten seeds from friends or relatives who are members of

COOPROSEM (Diakite and Diarra, 2000).

Willingness to pay assessment

The willingness to pay study showed that farmers were prepared to pay up to
about 200 FCFA for a kilo of sorghum seeds (Table 4-3), about 30% higher than the price
for sorghum grain. This is a surprisingly high price of seed given the low income level
(CIMMYT, 2007), the limited cultural concepts of paying for seed of a crop for which
the markets are weakly developed, and a setting in which seed is exchanged generally as
a gift or through non monetary exchange (Siart, 2008, Diakite, 2004 and Diakite and
Diarra, 2000). The questions still remain regarding: a) Are these prices high enough to
create an incentive for seed companies or other traders and agricultural input dealers to

invest in producing sorghum seed? b) Is the quantity of sorghum seed demand sufficient

145

to encourage seed production of new sorghum varieties? The latter question was asked in
the survey, but clear answers were not forthcoming (author’s unpublished data), as
farmers indicated that the quantity of seed they are willing to purchase will vary greatly
from year to year. It is site and time specific, as it will depend on other avenues of seed
access, and the area that the farmer is interested in planting in a specific year.

Information on the cost of production of improved sorghum seeds will help us
address the first question. Diakite and Diarra (2000) in their study on seed sector
development in Sahelian countries, estimated the cost of sorghum seed production as
follows for Mali:

Table 4-15 Estimate cost of production of 1 kg of improved sorghum seeds

 

 

Designations Cost in FCFA/kg Percentage in total cost

Inputs 44 39
Labor 28 24

Motorized equipment 1 1 8
Amortization Animal traction 16 l 3

Small equipment 4 0

Quality control fees 6 3
Storage fees 16 13
Total cost in 2000 125 100
Actualized cost in 2007 145
Marketing cost (transportation and other 20
fees)
Total production and marketing cost 165

 

The cost used here was adjusted from 2000 estimation by using the Consumer Price Index (CPI) for Mali
from http:t/‘data.unorgtC‘ountryProfile.aspx‘.’erName—Mali, and an estimated marketing cost from “Office
du Marche Agricole” (OMA). Thus the costs in 2007 would be roughly 125*] .16 + 20 = l45+20 = 165 F
per kg minimum. .

Source: adapted from Lamissa Diakite and Alpha Macki Diarra, 2000

 

From table 4-15, the total average cost of production of one kg of sorghum seeds
was estimated to be 165 F CF A. Note these costs do not include research fees. If we

consider research as public good, i.e. no direct cost for plant breeders charged through

146

market in Mali, this total cost compared to the average price of 200 FC F A/kg farmers
were willing pay. If research has a cost this total will need to be re-estimated and from
there new comparison will be made between cost and willingness pay price.

Considering the first case scenario where research is a public good:

Table4-l6: Estimated revenue for sorghum seed production and distribution per zone
when research costs are not paid for through market receipts

 

 

Unit : FCFA/Kg Dioila n= 141 Maude n= 89
Average price (P) 176 244
Total production cost (TC) 165 165
Difference P- TC 51 =50 1 19 =120

 

The cost used here was adjusted from 2000 estimation by using the Consumer Price Index (CPI) for Mali
from http:l/data.unorg/C‘ountryPrOfile.aspx'.’crName= Mali, and an estimated marketing cost from “Office
du Marche Agricole” (OMA). Thus the costs in 2007 would be roughly 125*l.l6 + 20 = l45+20 = 165 F
per kg minimum. .

To summarize, if farmers were not obligated to pay for the cost of the underlying
research directly through the market, then the price that farmers are willing to pay would
exceed the private costs of producing these varieties with preferred variety traits. These
results confirmed the statement of Weltzien et al., (2006) that Malian sorghum growers
are interested in trying new sorghum varieties that combine their preferred variety traits.
However, price alone is not enough to start promoting seed sale. One needs to know the
quantity of seed demanded for these new sorghum varieties and the market for sorghum
grain (the output) in Mali.

In a previous study conducted in this region, Siart, (2008) found that improved
variety seed production and dissemination from 2003 to 2006 increased from 400 to
4,300 kg of seed. This survey indicates that larger quantities of improved variety seed is
produced and sold in Dioila than in Maude, which may be due to the larger area

associated with the Dioila zone. It is of concern that quantities sold in Dioila were

147

decreasing over time. In 2004, ICRISAT found that this may be due to recycling of seed,
so after initial purchases of new varieties farmers would not necessarily buy seed
(ICRISAT, 2004). Farmers also give seeds of new varieties to relatives and friends
(Diakite, 2004), which reduces market demand for such seed. This makes investment in
seeds of self-pollinated crops that can be recycled, such as sorghum, risky for production

by seed companies (Tripp, 2000).

Seed fair evaluation

Seed fairs have been used in Peru, Zimbabwe and Kenya to increase awareness of
diversity in crops and varieties, and enhance farmer access to a diversity of species
(Almekinders and Louwaars, 1999). The Mali seed fairs reported on in this study were
appreciated by participants (table 4-13): more than 80% of the 131 participants
interviewed were happy with the fairs and requested they be continued in the future. The
sellers at the fairs presented 19 varieties of sorghum, rice, groundnut, cowpea and maize
(Table 4-8). This is less biodiversity than seed fairs held in 2005, when 42 varieties were
available for purchase, and other crops were available as well such peas, millet, and fonio
(Siart, 2008) However, the number of buyers and the villages where the participants came
from were higher in 2007 compared to 2005.

The main objective of the fairs was to improve the diffusion of new sorghum
varieties through sales and to spread information about the varieties (Siart, 2008; Diakite,
2004, and CTDT, 2006). This appears to have been achieved, as our survey showed that
50% of the seed sellers were motivated to participate in the fairs as a means to bring more
varieties to farmers and to learn more about new varieties and seed production. These are

key elements in improving access to new varieties and the overall seed supply channel

148

(Sperling et al., 2003). More than 70% of the buyers said seeds in the fairs were good
quality (compared to the other seed sellers’ seeds). Some 35% of participants found the
price of seed to be above their purchasing power, against 42% who thought the price was
acceptable. The proportion of buyers who complained about the price in 2007 was higher
than that of 2005. In summary, the goals for the seed fairs (COOPROSEM, ACOD,
AOPP and ICRISAT) were being met, based on our survey of seed sellers, buyers and
visitors. As Diakite (2004) pointed out, diffusion of improved varieties depends on
information being made available about varietal characteristics, and almost all the
participants surveyed pointed out the need for more information on research activities in
their areas in general and particularly regarding traits of associated with new varieties.
Promoting information about variety traits and seeds available at the fairs will help
ensure successfirl seed fairs in the future, and increase awareness about new varieties

adapted to farmer preferences, and local agro-socio-economic conditions.

149

CONCLUSION

One of the important findings from this study was that farmers in both zones were
ready to pay 200 FCFA/kg for seed for their preferred varieties. This is a surprisingly
high investment given the low income level and the limited cultural concepts of paying
for subsistence crop seed. Generally, sorghum seed is exchanged as a gift or by non-
monetary exchange, as documented in this case. However, there was evidence that
farmers are willing to purchase seed at a price that would support moderate-priced
privately produced seed, assuming that the research costs were supported by donors or
general tax revenues. This suggests that there may be modest incentives for seed
companies or other seed sellers (OPCs in Dioila and COOPROSEM in Maude) to invest
in sorghum seed production, particularly in Maude, where farmers’ willingness to pay
was higher. However, it was not possible to document the quantity of seed farmers were
willing to pay for, as farmers considered this to be too site specific and to vary from year
to year. Seed recycling, exchange and gift of seeds among farmers indicates that demand
for large quantities of seed will be rare or non- existent. This uncertainty would reduce
the incentives to enter into private seed production in Mali.

Seed fairs were shown to be an effective means to promote local crops and farmer
access to diverse varieties. Farmers were happy with their participation in the seed fairs
and were enthusiastic about future fairs. Constructive suggestions were documented to
improve future seed fairs such as:

0 Promote a wide diversity of crops as well as varieties and widely provide
information on the fairs and the products to be presented in them using all

available means.

150

 

0 Have the fairs in many villages to make their access easy to many people and
have some fixed and known sale places of seeds after the fair.
However the overall quantity of seed sold was extremely low, and does not really
warrant investment of time and funds in seed fairs. The quantity sold is in no proportion

to the value of the seed, or the commercial gains the seed sellers can achieve.

151

APPENDIX 1

APPENDIX

Table 1-26 shows Number of seed samples collected per village and per crop type for the

seed quality analyses in the laboratory on 2007 in Dioila and Maude zones

 

 

Places Groundnut Sorghum Total
Dioila Seribila 23 17 25 64
Mangnambougou 6 l 2 21 3 8
Wakoro 20 4 40 64
Wobougou 25 1 1 18 54
Total Dioila 74 44 104 222
Mandé Gonsolo 35 0 23 59
Kenioro 8 0 29 37
Siranikoro 20 2 2 1 44
Siby 26 15 24 65
Total Mandé 89 17 97 203
All sample Dioila + Maude 163 61 201 425

 

152

APPENDIX 2

Table 1-27 shows the sample size of head of household interviewed during the survey
for seed quality assessment on 2007 study per zone

 

 

Magnabougou Seribila Wakoro Wobougou
Dioila Men 20 20 19 19
n=80 Women 0 1 1 0
Gonsolo Keniero Siby Siranikoro
Maude Women 17* 18** 20 19**
n=74 Men 0 0 0 0

 

**In Keniero the number is 18 instead of 20 because one head of family passed away and
one move to another place between the seed collection on the survey. In Siranikoro, the
twentieth head of household just refused to participate in the survey part.

*In Gonsolo, we could not find the three other head of household for the survey.

153

APPENDIX 3: Questionnaire for seed fairs survey

Name of the enumerator ............... # of questionnaire ................. Date

For seed sellers/ member of the cooperative

Name of Respondent: ....................... Have you ever participate in ICRISAT trial. .
Gender of Respondent: ...................... Village of provenance .............................

Q1: What species and varieties of seeds are you selling? Specify if it is local or improved
varieties
um Millet Groundnut Maize others

Varieties

 

For each crop notify if the seed is self-produced =S, or from others (friend or relatives
who could not attend the fair) = F

Q3: Why did you decide to sell these varieties?

Q4: Have you ever participated in a seed fair before? Yes .......... No ........
If Yes: as seller .................... or buyer ................

Q5: What are your motivations in joining this seed fair?

Q6: Are you satisfied with the seed fair?

Yes ..................... No

Q7: If No, What are the constraints to overcomes

If yes: what are perspectives (strenghs) to enforce?

Q8: Which other possibilities of effective seed selling can you think of?

For seed Buyers

Name of Respondent: ....................... Gender of Respondent: ......................
Village: ........................

Have you ever grown an experiment trial for ICRISAT?

Q1: How did you hear about the seed faire?

 

Other Radio Farmers Market Extension Invitation Others
farmers association workers/workshop By fiiend sources

 

 

 

 

 

 

 

 

 

 

Q2: Why have you decided to buy your seeds in the seed fair?

 

Lack of seeds Interest for Seed quality Varietal others
new varieties characteristics

 

 

 

 

 

 

 

 

    

Q3: What varieties did you buy and how many kilos? What do you think about the seeds:
' , 'ce?
Varieties Qty uali Prices

K VG G ' VB VL H VH

154

Note: Qty = quantity, Kg= kilogram, VG: Very good, G= Good, B= Bad, VB= Very bad,
VL= Very low, L= low, H= High, VH= very high For each specie, specify the name,
PM= pearl Millet, M= maize, S= sorghum, G= groundnut

Q4: Did you get all the all the varieties you were looking for here?

Yes No If no what are the missing varieties of which species?

How much of them did you need? ................. (# of kg)

Q5: What do you think about seed presentation (size of bag/label) ?

Q6: Was the information you received helpful in making your choice for varieties?
Yes No If No, What further information would you like to have?

Q6: What would you like to be improved about the seed fair to meet correctly your
needs?

For visitors

 

Name of Respondent: ....................... Participant in ICRISAT trails Yes. . .. No. ..
Gender of Respondent: ...................... Village: ...........................................
Q1: Why have you decided to come in the seed fair?
To meet others farmers ............ To relax ............... To learn about news
varieties ............ Just to look at ............... Others to specify .........................
Q2: How did you hear about the seed fair?
Other Radio Farmers Market Extension Invitation Others
farmers association/OP workers/workshop By fiiend sources

 

 

 

 

 

 

 

 

 

 

Q2: Did you ever buy seeds in any seed fairs? If yes, What (species) did you buy ......

 

 

 

 

 

 

What varieties ........... Number of time ......... Where ......
Q3: Why didn’t you buy anything today? Prices too high ........... Quality not
attracting ............. Don’t know the varieties
Don’t have money ........... Will buy later ............. (where?) .......... Already have
seeds ............. Others reason to specify
Q4: By looking around what do think about the fair?
In general Place Day in year Week day
Very good
Good
Bad
Very bad

 

 

 

 

 

 

Q5: What are the possible improvement areas for you?

155

APPEN DEX 4: Seed samples collection surveys questionnaire

Number of questionnaire .....................
1- Identification of seed source

Zone ..................... County .................. Village: ..................... date ................
Name of the head of family: ................................. Name of the person interviewed:
Crop Type .............................. Variety Name ..............................

1- Information on the seed lot

 

 

Year Seed Seed production methods
source
Variety Off type Selectio storage Preparatio
planted in elimination n of u of seed
isolation before panicles for
(individually) flowering planting
or not

 

 

 

 

 

 

 

 

 

 

156

APPENDIX 5 : Questionnaire for variety evaluation 2007 in Dioila and Maude
zone

Name of the enumerator: ......................................... date

Name of the head of the household .............................. Gender ......................
Village ................................ County ....................

1- Have you ever done an ICRISAT trial ............... Yes=] , No=2, others (to specify) =3
If yes how many times ............... Which trial .................

2- After done with your evaluation we want you to give us your impressions on your
three best choices of varieties .............

 

Name of the variety Reason of your choice and ranking

 

]-

 

2-

 

3-

 

 

 

 

3- If the seed of these three varieties were available for sale on the market, how
much will you be willing to pay for the kilo if the consumption grains cost more
than 150 Fcfa ? The top best ...... Fcfa , Second best ...... Fcfa, 3rd ...... Fcfa

4- Have ever planted one of the above listed varieties? ........... Yes= 1, No: 2, If yes
which one ................

157

APPENDIX 6: Head of the household interview questionnaire

Zone ......................... Commune ........................ Household number .........
Village: ..................

Name of Respondent: ................................. Gender ...................... Age

group .................... Young= 1, Middle age= 2, Senior (old) = 3 Are you
responsible for farm work (chef des travaux)? . Yes= 1, No =2 , If yes continue the

interview with him/her alone, if no ask for the person in charge of farm work to come and
associate him to the interview

1- How many people are in your UPA? .......... (Number of people eating in your UPA);
How many work in family field firll time ..............

2- Is any of them involve in other group/collective farming activities such as varieties
trials, farmers association, etc..?

3- What is your family fields’ size? (# Hectares) ..........

4- What do you grow in your family field. . ...? 1= sorghum, 2= millet, 3=millet and
Sorghum, 4= sorghum and groundnut, 5= millet and groundnut, 6= all three.

5- How many women and men in your family have their individual field?

6- Do you know what crop they plant on their field? ............ Yes= 1, No= 2

If yes how many of them are growing: Groundnut ......... Sorghum .............
Millet ..........

7- Do you know how/where they get their seeds? ..........

8- Level of equipment .............. (Welll7 =1, Average]8 = 2, Not'9 =3)

 

'7 Well in equipment means the farmer has all common rural farming equipments used in Mali : plow
(mold-board, gang, disk, lister, cattle (ox), cart (donkey cart)

'8 Average: the farmer has only one plow and one pair of ox

'9 Not= no Equipment

158

 

 

208008 238.0 0080: 3:080: 0: 8320308:
3 3:3: -080 03:08»; 20 8520808: 30588 0: 0820308: 6:0: 0:43 0.20 0880800 0:802

 

 

 

 

 

 

 

 

000:0800 88> :0 N0>0 80> 0:: 00 “m0 :0

30:00... 3.58 0.8.: 0:68-033 .5283 :8: 80.80? 880800 02 _m .V

$08,080:: 20 E ..................... 00:9 30 08:3 00 080: 00 00:82: 0:08 a N :m

m ”80:20.08 3 0880 .N u0w2=> 08:0 ._H0w2=> 0808 ”080800 2 0: 00:0:0>8: 20 :0: 2 So: am

.C0E0.0:m 3 0083 8:283 .0880: 0:50 .280 .829 A883 2:80: 2 0: 8:822 ” N. 08088 88> 82:0 N0>0 80> 8: 0Q
.0880 8:00 080:3 502000 ”0080800 020 82:05:00.: 2008 2 80:

......... 0088 080:0 20 82:0 w: 20 8:80: 2 N088: 0:88: 2 :5:

............. 0:0 0:80:08 20 88:58 .680 .280 £80.88 00:08: 0:808 8200808: 8:8: ”080800 2 0: 0838

m ”£08008 3 00800 .N ”2800 020 80808 :0 43:00.2 2 08:20 8082800.: 0:82 2 80:

 

 

 

 

388080: 82:08:00.: 080800 22...;
0: :21— 8: 0Q 0602 82:08:00.: 0:80.: 838:0 2 0: 3.80m 2 0: 80 Z

 

 

 

 

 

 

 

 

 

“3cm 0300 022:0 00>: 25> 0:: 02:080.: 022:; m2 N000: -m

v n $0208: 8 008:0 hm H0808 :08 .N “00:08:80 08020: 020 :25 4 ”0:80: 20 8:0

N. :80:0 00 20 080800 :0 80808208880: 80: 82200: 2 88: 80A

088% 00 :8: 00:08:80 8028 020 :0:0 2 :o 0808 30 .«0:0 2 N28080: 82820800.: :0 02080800 20 :80:0v

088» 20 83:0 8 0:20 :0 00 2:80.: 20 :0:0 80 882:: 000.0 :0 00:38:80 03028 020 8:0 2 02080820 42:80: 88:0 2 0:20 5
v "C0208: 3 08:: .m H 0808

20 88:0 .N ”—08:28: :880 4 ”2:80.: 0: :80:O:::.::.e .0280 0:3 03:00.8 0.:o> .082: 28>- 00>: :880 .3: Em A

omAI I 82> 0m 9 mm ”:088 08 mm v H082. .................... 0w: :0 0880

G~cneoow “N nSfisomdanw OED” 0~ SOL ......... ..UEOG ........................................ ngccoaoh :@ DfifidEflrw Omvawofl—U at 802
......... ............DEOO 2...................................... .chwcogm zwgoz

.............. ...e............O$§fi<mb..........................-.Qwa~fim>.- ......................QC:EEOU .........................DF~ON

303:0. 9.039% 83888 0:58:80: 03:00.8.— .50: 153818 02:80:02.0 " b X.GZH:.—<

159

%:0> 80: 08:080.: 3583 0: 0: 050:: 0: 0::: 200:0 000:0: 0:000:20 :0:0::8: 0: 008:8: m0: 0:080: - : .0.w

 

u 0w008000o 20:80.: 8:00:80

:80:0 :: 8:000:00:
0w0mm:0wm 808800 ” 0:3
80: 080—: m0: :0:088:0

 

0w0:00m 0w0_80m 8:202
::0:O

0:000: 280m 80:08: :80:0 :: 80:0

0:000: 0.080800 0: 8200200 3:980:88: 00230:: :88? :: :0::_:0m0n

 

 

 

 

 

08083 0: :0_00::0._: .0.m

0:028:00: 3:00: 38> 0: 8080:: 080800 0- 5 -m

80:808 :: 0080800 02 0000:080 808088: 80: :0: 080 80: 8:0 N080: 80> 0:: 00 $0.30 .:08m
Nu:0Z .TSo ................ m. 000:0:00 N0>0 m:0> 0:: 0080800 02 00>0 8:800 m:0>-m0:m

00:00.: 08:00 0: 0: 80:0 :: 8020: 02 80 8:00:80: 02: 0080: 80: 0080: 80> 0:: 00:0:

 

 

 

 

 

 

 

 

 

 

 

82:8 0m0x0oum 0:
C8608 0:803: 20:? 0_ 80::0> 08083: :0200 0080800 008038: 0: 00:00
0: 00:30 0: 0: 0000020800 :: 8:000: 800 0: 0: 8:088 00: 080:0 0 0000800 0: 0:0:

 

”$030880 000:0:0: 02 02:3 00:00:: 40:00.: 08:8 0:: 008:: 83.030 8880: I :6

m N 808008 8 008:0 .vu 9:000 :0 2:0: 0:: :0 808800

:0 0:00: 0:3 030: m2 .mn080: :0 ::0:0 .N H 8080 :0 050:0: ::0:0 2 “800:800 8: ...........~_0:0:00 N0>0 80> 808800 0.0.:

m H A0800: 8 008:0 .0 ”0220008 02 . m “080% 0: 80:80: 02 .N

“$0880.: 02 80820.0:0w 080::0> 8:03 0:808 :0 08:080.: 00808 0: 080::0> m2 .2 ”08:000.: 8:20.88: 00: :0:0_0omm<
”08:: N0>0 80> 0:: 080800 0: 080:0? 0: 00:3 02 020: «.0:

......... .. 080:0? 080> 0: 0088: 80> 80> 00:05:: 0:08 <

.........................:....00:080m 0: 80::0> 88> 80:0 N0>0 80> 808800 :06

 

160

000.58 00. 08:3 0:800:00 320: 0300 800:0800 m0: 058:. a. 0: 553003.; .5

 

 

 

 

 

 

C0560“? 3 m0bs< m0u0t0> 8:025: 00 86:55 0380:? £80m 00:0E0m 00 8300:8232

 

20:: So: «:280 $0588 000 6:80:00: 00 5:355 0::.0 m00:0E0m .000 :0 N02: 96> _m .0
05:30:: 02 85> 808800 .0:
3300.6 :0 05:05:08 N0>0 mso> 03¢ 8:088 00: 05:00 805800 mo> 8&0on :5: m=o> 5:08: 85:80:: 00 m0t0m 820:0 0-4m

 

 

 

 

 

C0E00mm 0V m0::< m0:¢_m%:mw0wm:::oa m080m£ 0:283: 05083809

 

 

 

 

 

c:0:w=:m0.m Sm 80:80.: 02 000:8 Nggmev 0000583 mo> 00 050::

2 00 50330008: 00 08:0: :0 :o::0:0 8:: 00:8 m:o> £03930 8:088 .02 “:8 £03 805800 000 0w0x08m 0: 0:8:Q -m-:.m
............. mmolw ow :DMQEOU ........................ ...................ޤ=O

mnC0Eo0mm 8 00:50 6 "682:3: 0:080:80 am ”02030: 0.630: .N H008 m2 0:00 02280850 03:8w 000 8:855

J ”8:338:00 00 008: 2 0:00 0220: 02200005 :o:8=&< ........................ .5520.“ m:o>-~0>0 0:080:80 00 093 6:0 “:5 _m
Nu:oZ a T50: ......... 000:0800 0:o> 00 0833on 0_ 88% 0:080:80 :: :8 m:o>-N0>< 95%

v u 503% e

00:30 .m “coca—8:88:00 00 025080; 00 E08038 :m “N ”02:83.0 00:00:; 00:20 000 0:080:33 A M 0:53 00:20.0 “5808:0m
................. 0008 000 N080w m:o> E08800 I m-n.m

v "£05009, 8 830 nm ”:88 :o :00 00 05a :0 008 .N ”0:08: :0 88 E0: 4 ”0:08: :0 0000 £88m 004

................... 08080 00.8 00 8:8 20:0 £08 m00 880 308% 0:880 02580.0 m0mmsow 00: mm I N.:.m

 

 

 

 

 

 

 

 

C0£00mm® m0bs< 03883 2 0:0: 000.0on 008 :m 82:83— :30 0003on 0000 :m 880% 0_ 0:8. 08> :m
0025890 m0mmsow mo> 28> N0>._0m:oo 808800 J mm

80203800: 850$ :0 Hm .m0mmsow :0 H: ............... 000.383,“. 8588 0bo> 33-88% 808800 I 79m
2000:0800 m0: $8.005 Rm

.................................. :0 am a :05800 80 5

. . . . 0058:0858: 0:80.: a 0::00 30> -N0>< ............ 00:25:: m:o>-N0>0 0:6 8:088 00 05:85 0:0:O-m.0.m
m ”£80m 0_ 8030 000:.“ .Nn0wmno0m 2 $80 . 70:80: 00890 003.0 . .. ........ 00588 8 00 080280000 0_ So:

.................. @DH—OUWH fl— WMHQN Bay—.0“ 0933800 «N am ,

.0280: 008% 83% 833010 Mm :830 :0 0:80.. 2 0 03:“ ..L H 0w8m:ow0._ So:

161

..Comawnm S 8E5. Q
.. ........... GOSNHQNU< A0
............... mvfimwa mOU SD—SOU Amav—kvz. . .. . . . . . . . .. . . . . . . . .uggvfiom RwMOE DU OHQEOGV... . . . . . . . . . . . . Zwmumuwww> 03th An
A... . . . . . . . .80 .8885 8c 2:8.“ 6:8: 082m 2 8 oEEonEoE Q
:2 32...; 088 on. 839:?» 8:3:8880 8- 8288.. :2.— mzo.» Fm A;
................... 828888 a 853 .2 8882 28 08:88 2 Bog 22:; 0:8 so 82“ 80> 8c 8 “8 28:0
....... 080C COW Hwo —ODO
$.88 038 2:28 88 80> 08 258.88 2.222 2 H8 2280 ad
"338; a. on. 8:83.?5 um

_ _ _ _

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

r Cowwommm 8 85st _ 2:3 2:59 8 82¢ 8 05:82 _ 8>2 2 a Scam; 8522::me 8 832—. _

@288 83m 80888 8w 2:25 2 8889 a 80> EoEEoU A5.

8.28% m

3 853< 85:8 858% 85me 88 888i 2:2 8:8 52:8 1
$882 8 N285“? 80> 28588 .89 Son 828%: 22m28> 283m 2 2m .mdfi
8888
€2.26me 8888.8 822m 2 8958 88988
3 825< 882?. 2:2 5258 m8 8:8 «2 28:92 2 8 2:2me5
$882 8 8 238 a 888m 2 H2888 80> 82:88 Sou 82.8 Em 8 888mg; _m .Ndfi

@8888 2 8 .2228 2
“2858 Son 528m 8 825: 28> «8888 88:95.8 80> 825m .89» Son 82.55282 #8 58:? av 83239.2 2 mm :5.

 

 

 

 

 

 

 

 

 

 

mflwug mBBoscoE
CoEommm m2 8% 888 2323523 520 828 8c 228> 58:8
8 823‘. 88898.2 8o 822585 2 :5 862582 .6 852382 888mm<
528m 8 80:8 2 Egg 882:8 8c 2:25 2 25> 882 an EoEEoO -afi

c H C0E008v 85:0 mu 2808: 0:82 8 ..cosamcmnfifim:

.v ”0:288: $828808 20000.83 .mn208 8:008 .2H208 8:88 ............ $208 m:0>-N0>0 :83: 8:: 2:0 :m
N n :02 ._u_:0 .................. 02000.— 8 8:0 m:0>-8fi

A8888: 8 858on : . . . . .088 80> 8 83: 8 :88 8 8:8 82:88: 8888: 8 85600

9002005.: 058on .. ............. :88 :8 2:5.

8.38.8 8:05.288:—

 

 

 

 

 

00=n8>0a :0E2:80% 8888 3.8:?
8 :0: 3: 0C 802 :2223858 3E3. 0::—«:0 2 8 8::0m 2 8 E02

 

 

 

 

 

 

 

 

@3003 088: 8:8; 8 888m 0bo> :850 m:0> 0:08:80 -0

 

 

€050QO 88850.8 028:8 888:: x8 858m 8

av mobs< 0280008 858253 2 8 008GB 8352.: .0888: 88883: E05088:

 

 

 

 

 

 

 

 

 

88:08: 80:00 8: H25. 08.53 080 8 .508 2 50: 8:: 80:8 80> _:dlm:208> 88880880 m2 20:0 .0. a
......................................................... ......... :owggom=o>l NQ>N~OSUH=OAH

..................... 020.: 08:88 2 80: 28588 208> 88> 888 N0>0 80> 0:: 8 80 08:0 -8. a
8:28 m3: 888 608588 08.9; 88> 80 208> 080 58:5: 6.:

163

2.00888 88> 8 88 80> 0:20 00 80 :0

80200.8 00.88 088.8 0.2—0-4.08.0 .888.» 8:8 80.82.» 80888 8_ 5 .8

8020822 08 05 . ........... . . . . . . . .. m88> =8 0w2=> 00 0802 8 008028 0:80 8 N mm

m ”£050.08 2 828 .N H0w2z> 880 .2H0w2:> 0808 ”00888 2 08 008888 8 :02 2 800 um

202.2008 2 82:0 8:888 .8088 880 .280 .828 28023 2:82 2 08 82822 n N. 00888 88> 8220 88 80> 88 0Q
.8200 .8200 .8828 .8800 800888 88 202228028 0808 2 802

......... 0088 0:820 08 0850 92 8 8280: 2 81288 .2288 2 80m

............. 000 .0280208 08 8028882 .808 .280 .2883 82008 82088 8020808 0808 ”00888 2 8 088m

m “282.50% 2 828 .N H288 88 80808 30 28200.8 2 2388 8028820028 882 2 802

 

 

 

 

0088080 :8 8222020028 00888 80208
8 =2.— 09 802 8222:0028 880E 02880 2 8 00.80@ 2 8 802

 

 

 

 

 

 

 

 

 

"Sow 0:00 .0228 88 80> 0:20 0238 08 8.2.2:; 8. 008: -m

v n C8608 3 828 .m H0808 88 .N ”8288020 8028 88 .2020 .2 H2282 08 .2020

................ m. @8020 00 08 00888 8 8080808308022 800 82208 2 888 SNYN

.0808 00 802 82808020 8082 88 .2020 2 :0 080% :8 .2020 2 888808 8082008828 8 03088000 08 @8023

080% 08 08020 8 88 8 00 2:82 08 .820 :8 882.28 80.0 8 8200820 8082 88 .2020 2 0088808 22:82 08020 2 808 2m
8 ”2002008 3 0.80 .m H 080%

8 @8020 .N “2082288 98820 .2 “2:82 8 8820:555. “.0880 0:00 0298 88> 082.. 83-88 28020 .050 8m -2

OmA H xsomx/ .0m 02 mm ”8%08 OWW .mm V 8685—. .................... Dwayv 0950.20
HHZHHEWM “N “CSSUWMEHM 0H8” O~ SCAM ........ ..DEDO ... ............. . ...................... .Hg—UGOQWH 3U Q=§% OUMOSU 3U EOZ
...................... DEOO ........................................ .Hgficogm 5680.2

............ ..................OHwEHZH<AHD...........................HOwN~—m>........................Q§EEOU .........................O:0N
8.2080585:

8:8 8030..» 1882288 02:8 18.— 88 8:29.02 :8 050m 8 0288 2 80m 8:82.288 088828050 " w VGA—Z???»

164

 

 

 

owwzam 28w5.w 82338? @883 8 82888.05
80888 H oEBwfi€ owwofioimnm
20 3829 wow 82888$ conflofl
owmnoww
owflewm
@830 80830:
0:08: £80m @820 8 50:0
833.. 82383 3. 858—3. 8:388:03: 89:an ASE—o :6 853882—

 

 

 

 

 

3583 on 859353 .a.m
“3588 on 85352.— oEoE 9:68-399 85259:— ouaoa «.59» on 823.:— 8588 2 _m -m

$5088 30 82588 m2 yoked—mam 3080808 Son 8a 26 8am Sc Newman 25> 0:: 8 “8.5: doEm
m. Boson N05 30> 25 $0583, $152“ “Emumm 33%on
$88.8 880m 2 cc 8080 8. 888 we“ :5 82282898 33 888 30: 588 93 25 00-8mm

 

 

ommxoofi ow
€2.08QO 8.08; ‘3ng E .5282, 3:283 808% 32583 882205 on 828
8 88% 88038 «A at 02836800 :8 808804 Sou 3 av 808255 mow .2830 m 858% on 8.5m

 

 

 

 

 

 

 

 

 

 

AmuEwozmmm 880.68 m2 888 N823 588% 888 on: N825: m:o>-No>a Bianca I 8v

v u Commowmm 8 883. .M.“ 938 so 2th 08: 6

25888 as 283 083 8% m2 .mnobdw: no :35 .N u 80mg“ 5 29¢me :38 4 H530 ...........~.Bonow N96 35> 808800 -m.a.v
m H 38.52. 8 3me .v ”836mb m2 . m ”28me no Efizfiwu m2 ,N

"@0800 m2 Bog—88% 8:28? 3303 £088 on 0280898 88me um 8528? m2 4 ”02:83.8 8532608 mow 8033008<
”3:3: No>m 98> 26 88088 20 8:28“; 20 890 m2 Noam: 98> Fm «.56

. . ..... . . ........ NHDOUGO> obo> DU No>505 m50> mdo> oogummfi 0:036 < ............................... woCOEOm

no 538? 88> 6580 N96 m=o> 808800 ..................................... $858 28> :88 N93 39» 808800 fiaé

manSméSN $88 8:083 ow 838268.“
coEownm 3 mob=< $0558 ow 8:226on mouaquw 3. 338355 AUmOV 0.53 mow 0883000 3832

 

 

 

 

 

 

 

 

 

 

 

165

 

 

£05080
8 00:844. 089003: 00w0880Q 0000008 0828:: 0:806:83.

 

 

 

 

 

 

 

C:0:w:mm0.m 8: 0080.00: 02 0088 3:80.80 0008088 00> 00 050::

0_ 0: 8830008 00 08:00 :0 :28050 0:8 00:0,: m:o> 283:0 $800000 m0. 800 20:: 0080800 00: 0w00—090 0_ 8085 -m-a.m
...... .......m_o;* 0U GOMDEOU ........................ ...................mU§—)—O

”C0060“; 00 00:80 ,0 1082800: 8080:08 .m ”03050: 08:0: .N ”0000 00. 0:0: 02030008.: 280% 00: 880058

._ “8:088:00 00 0008 0_ 0:00 02:0: 0203088 8:00:83 ............................. 0:00 m:o>-N0>0 80:80:05 00 0:5 6:0 .80 _m
mn:oZ ;uSo ........... 0080800 28> 00 0m0x080 0_ 8080 808080: 8: :00 m:o>-N0>< 0.36

m H 92.0680 00 00:80 .N “05805 0_ 80: ._ H 080% 0_ 0:00 ......... . ....... 00000 000 N080m m:o> 808800 I m-a.m

v MAE—050% 00 0:80 m ”:080 :o :00 00 05a :0 0000 N ”05008 :0 000m 80: H ”0500—: :0 000m 080% m0:

................. 080—000 0000 00 00800 00:: 0000 000 0:0: 808% 820:0 008080: 00: 5| N. a. m

 

 

 

 

 

 

 

 

 

0080:»
0_ 0:0 8:050 000390 580% 2 0:00 00.00330 05:0 0080:» 080$
2:080: 00 080w :m 0080.80: 00 080w :m :: ::m 0:80:80 030» :w 0_ 0:00 0.3020 0000 :m 0. 0:00 08> :m
mowsm 0: 008080: 00> m:o> 5300800 808800 J _m
C08008 8 0:80 Hm £00883 0080a :0 MN £08080: :0 H_ ............... 093:8 00 080800 0bo> m:o>-N080w 808800 I 79m
0000:0800 00: 093—005 8m
.................................. do Hm 0 805800 80 mm
. . . . 00888208008 880.0 0 8:00 m:o> -N0>< ....... 008808 m:o>-N0>0 080800 00 00:80:: 0:0:O-m.0.m

mm:o> 8o: onwsm 00 080800 0: 00 050:: 0_ 0::: 0000.000 00000: 03000-6 80:00:00.8 00 00:88:: 000 0:0:004 -v.0.m

N “0800 0_ 80>0 000:.“ ._H0m0:000 0: 0080 ..:.:.:: 0080800 0_ 00 0080800 0: 0: 08:0: 0_ 00:00 w:o> 0:: 00 :00 80:0 -m .0.m
............ 2....myou—OO‘DH M# WWHQM so“ 0“ ”CmDEOO «N mm

.0280: 0080 8:0“ 8:308 MN @8080 :0 0:000: 0_ 0 000% H:

...... . 00830: 0: 00:0,: m:o> 0:: 00 :00 80:0 -.m.0.m

.. . . . . . . .. $00000? 2008 :0 0w000 #0::

< . . . . . . . . . . . . . ........... wQEmSo SM 300 mow—flaw m30> mmom 20 62388 Dughwhmv Damofioflafim 0U mug—Q mow GOMU—Mfimg—wL Son :5 —m ufldfi

166

 

 

 

€080.80 0V 008:0

 

0800 0800 80 80:30 0880 Z

 

00>0_ 0_ 0 0:0:w_>

00008808 00 0300.

 

 

 

00800 00000 00000800 000 080:0 0_ 006000 0 0:0> 808800 -06

 

 

 

C080000 0v 00b:<

 

00.880 00080.0 00808 000 00800000

 

0800

 

 

080.00% 030—300

 

 

0800000 00 8600000 0:0> 808800 .0:0> 0:00 8080080 0_000c0> 0000800— 5 0.0.0

 

 

$080000
00 008:<

 

 

00000080
00:00.80~

 

 

0800 82:00

 

000 08.80 04

00808

 

0008080

00000800
2 u§>=oo 05009805
2 00 008008

 

 

 

0800000 00 00 .8000 0 0000800 0_ 00600000 0:0> 808800 0:00 8080 80 00 0080000000 E 80.0.

00080800 00 00 080:0 0—

000000000 .800 80800 00 00080: 0:0> 808.800 00:00:05.8 0:0> 00>:00 .0:0> 800 8080080 000 0:00:0> :0 80008000 0_ _m 00.0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0000t0> 0808:0000
£08080 00. 00>0 000000 0800008038 520 008:0 000 00080> 80080>
0V 008:< 080800000.)— 000 80808008 0_ 0:0 800080.80 :0 00008000 0080000040
000800 80>0 00000800 000 080:0 00 80> 000000 00 808800 0.0
000800 00- 0083 080800 00800 0300 "0080800 000 080:: 0. 00 80380000000. 4.
8050000 3 0008.0 0000t0> 00=0>:0: 00 8000805 050008? 8000mm, 0080800 00 8000080822

 

 

 

80:0 800 800000 00080800 000 6080008000 00 888808 08.0 00080800 000 :0 N0>0 0:0> 0m -0

00083000 000 0:0> 808800

.08 80000-6 80 080808 N0>0 83 0:0 080008 02 8800 00> 080800 800 83 N808 82800000 00 00800 080:0 0-8m

167

fiBSooFU 058076 ............... 95:0 2U mink
3.22% Earn—58:—

 

 

 

 

35396.3

 

 

 

2. :31— m.... on—

55:33."...

 

 

«to: 553.33% man—ob 3:530

 

3553
a. at 33cm

 

3&2;
2 3. =52

 

 

mAcooflowmmam $58.— 3 3:083 88> 2:050 30> E08800 -a

 

 

Cofiowmm
3 mobs<

 

 

ohgfionvow $03323:

 

0322—03 28515
2 on ooaflfloh 0955mm

 

 

$885 53
3808.5 8.8363“

 

 

mofifim co
EoEowcom

 

 

”$38me 832 No: .225 .mfitg 0:8 ow x85 2 Son 33 Eggs 30> Eu xsfiwwaxz 3:33.538an R: 2030 ..o.»

........................ oco.o..o.coco-c.0ooooooooooooo.coo-o.oo.oouo......\ow:w£o wao>lND>m wosv “SOL
..................... ~28 Ewing 2 Son 3.358% 32.5» 28> mwSEo No>m 25> 26 8 “mo 28:0 id

”mzflmc 33 Moscow dEaEEou Emu? 88> “mo ‘32“? 0:8 823:8 -ud

”€059on 3 9825. E
.............. COmHNHQNU< Aw
............... mDmeHw WOU MSU—DOD «Awgvfiwv... . . . . . . . . . . . . . . . . . . . .uCOEOmvfi—Qm ”AwMOE DU DHDEOCV... . . . . . . . . . . . . . . ..:HNHWWWN» ®~O~AU C

A ........... Bo £05285 mow 258 6:83 2.83 2 on 2w2on9o2 Au
:3 33...; 8.3 2. «333...; 352223933 8- $3.82. 3:..— m=c> :.m ..:.w

.......................... £88308: 6 028m.— NoEsmov flow 8&an 2 Son 33.5» 380 so ~o>m m=o> 25 8 “mo E30
.................................................. AWEOC com um” —o:o

wows.“ 0:8 @323 No>m m=o> 25 85588 33:2, 2 «mo 0:050 ad
“33...; 2 at 3.8.6.3.:— -m

168

o H Commommmv .853 mu 2223: 2&5: cu ..cogmtwnwafims

“v ”2663:: .muotmwcooom 2832.63 .NHBQG 288m gnu—98 >2an ............ $528 m=o>-No>m :83: 6:6 Bo _m
N u :02 JHSO .................. ““28?— m 8% 39¢on

$228qu ow mounEoZv .......... moEmE 25> at 33 :0 9:20 00 22% Beauty mogofioq ow 556on

169

APPENDIX 9 : Grouping of the farmers preferred variety traits recorded during

the on-farm variety evaluation.

 

 

 

Original list of variety trait Grouping
Long panicle with a lot grain, long big and Yield

heavy panicles, high yield,

Early maturity, this variety is more mature Early maturity

than the other planted at the same time

 

Plant with green leaves, strong plant on striga
infested places

Drought and Striga tolerance

 

Color, size, beautiful

Grain appearance

 

No matter when you plant this variety here
you will get something, adapted to our
conditions, strong stem,

Adaptation to the cultivation area

 

 

Anything else than what are listed above

 

other

 

170

 

APPEDIX 10 : Liste des codes

Genre: l= masculine, 2= fe'minin
Group d’age : 1 = J eune (moins de 35 ans),
2 = age moyen (35 to 50),
3= Vieux (plus de 50 ans)
Type de champ: 1= Champ de famille, 2= Champ individuel, 3= Champ de groupe, 4=
autres
Decision pour l’approvisionnement en semence du champ :
l = Chef de famille
2 = Chef des travaux champétres, 3 = moi méme, 4 = autres (a préciser)
Nom de la variété : Pas de code
Source de la semence : l = propre production,
2 = marché,
3 = proches (parents, amis, etc.),
4 = institution de recherche,
5 = autres (a specifier)
Quantité : 1e nombre de kg obtenu de chaque source
Temps d’acquisition : 1 = depuis la récolte,
2 = au moment des semis,
3 = autres (a specifier)
Mode d’acquisition : 1= propre production, 2 = cadeau, 3 =échange, 4 = achat, 5 =
autres (a specifier)
De qui : l= Moi-meme, 2 = Membre de la famille (époux), 3 = voisin,
4 = amis, 5 = autre parents (oncle, tante, belle famille, etc..),
6 = vendeurs, 7 = autres (a specifier)
Lieu de provenance : l = méme village, 2 = autre village, 3 = autres (a specifier)
Semence de source d’achat
Place d’achat : l = Marche
2 = Cooperative des producteurs de semence
3 = ULPC (OPC)
4 = Producteurs de semences formés
5 = Producteurs de semences auto-établis
6 = Autres (a specifier)
Maniére de choix du vendeur : 1= proximité
2 = Lien de parenté
3 = Information a radio
4 = Information aux marché
5 = Information des autres producteurs
6 = service de vulgarisation/ONGS
7 = autres (a specifier)
Distance au vendeur : Nombre de kilometres a faire pour arriver au vendeur
Type de vendeur : 1 = Association des producteurs d’arachide,
2 = les vendeurs de graines d’arachide au marché (petits vendeurs
généralement les femmes),
3 = les détaillant de grains,

171

4 = les grossistes,
5 = autres (a décrire)
Mode de payement : l = Au comptant,
2 = credit payable en argent,
3 = crédit en nature,
4 = les deux (une partie au comptant et une partie en credit),
5 = autres (a specifier)
Raison du choix de la source d’achat : l = Perte de semence a cause de problemes de
stockage
2 = Qualité des semences
3 = Situation de la saison pluvieuse
4 = Location du vendeur
5 = Connaissance de la variété
6 = La diversité variétale
7 = cout
8 = Autres (a specifier)
Satisfaction avec la source : 1 = Oui, 2 = Non
Semence de propre production
Description du champ : Choix du champ (rotation) : 1 = Oui, 2 = Non
Isolation : 1 = Oui, 2 = Non
Preparation du champ : l = nettoyage, cassé les mottes, et
labour
2 = nettoyage et cassé les mottes
3 = nettoyage et labour
4 = nettoyage simple
5 = autres
Practices agronomiques : semis : l= semis sur terrain labouré
2 = semis sur terrain non labouré
3 = semis sur terrain sec
4 = aures
Sarclage : Nombre
Mode : 1: main
2= multiculteur (charrue)
3 =1es deux a la fois

Elimination des hors types : l= oui

2 = Non
Application d’engrains : l= oui
2 = Non

Sélection de semences : 1= oui
2 = non
Quand : 1= au moment (16 la récolte
2 = au moment du battage
. 3 = autres
Comment : 1 = Sélection panicule
2 = Triage de graine

172

 

 

3 = autres
Récolte : Récolte : l = sur pieds, 2 = tiges cassées,3 = tiges coupées,
Séchage : l= au champ, 2 = a la maison
Battage : l = a la main, 2 = Charrette, 3 = Machine
L’egoussage : 1= juste a la récolte au champ
2= quelques jours apres récolte
Moment l’egoussage de la semence d’arachide: 1= juste apres récolte,
2 = apres 1e Séchage,
3 = juste avant le semis
Moment du battage de la semence de sorgho : 1= juste a la récolte au champ
2= quelques jours apres récolte
Quantité de semence produite en Kilo : nombre de kilo
Autre producteur/trice ayant recu la semence : l = oui, 2 = non
Combien : nombre de producteurs
Stockage des semences : 1: en panicules,
2: en graines vannées,
3= autre (a préciser)
Maniére de stockage des panicules : l = En vrac dans le grenier,
2 = En sacs stockes dans le grenier,
3: En gerbe suspendue sur un arbre,
4 = En gerbe de panicules stockés dans le grenier
5 = En gerbe de panicule stocké ailleurs que le
grenier
Type de sacs pour garder la semence : 1 = Les grands sacs en plastic,
2 = petit sacs en plastic,
3 = sacs en fibre de dah ou coton,
4 = autre (a specifier)
Maniére de garder les sacs : l = Dans le grenier,
2 = dans la Chambre,
3 = autres (a specifier)
Traitement des semences : 1 = Oui, 2 = non
Types de traitements : l= Application insecticide liquide dans la place de conservation,
2= Utilisation des granule d’insecticide dans les sacs,
3 = poudre pesticide
4 = traitement traditionnel,
5 = autres (a specifier)
Reste des codes
5.b-4 Quand : l= au début du stockage
2 = au milieu du stockage
3= autre (a specifier)
5.b.5 facteur plus préoccupante durant le stockage :
1= insectes attaques
2= humidite'
3= Temperature
4: dommage physiques
5= autres (a specifier)

173

5.b-6 : dispositions pour prévenir ce facteur :
l= garde la semence en gerbe ou gousse (pour l’arachide)
2= Traitements
3= Surveillance de la semence durant le stockage
4= Autre (a specifier)
6— si la semence provient d’une institution de recherche quelle était la raison :
1= Multiplication de semence
2= Essais variétaux
3= Diffusion de nouvelles variétés
4= autres (a specifier)
7- Appreciation de la qualité des semences :
7.1- Appreciation avant semis : 1= apparence
2= Reputation du vendeur
3= Information de autres producteurs
4: Information des ONG/vulgarisateurs
S= propre experience avec la variété
6= Autre (a specifier)
7.a.1 — Reputation important : l= oui, 2= Non
Si oui quels les facteurs qui vous orientent :
1= Serieux, 2= Qualite du produit, 3= prix de la semence, 4= echos des autres
producteurs
7.a.2- L’apparence important : l= oui, 2= Non
Si oui comment vous apprécier l’apparence :
1= Intégralité de la membrane couvant la semence
2= pre'sence d’impureté
3= uniformité des graines
4= uniformité de la couleur des graines
5: la dimension des graines
6= Trace d’attaques d’insectes
7= autres (a spe'cifier)
7.a.3- La pureté variétale important : l= oui, 2: Non
Si oui comment vous appréciez ce facteur :
1= uniformité de la couleur
2= forme des graines
3= dimension des graines
4: presence des graines d’autres cultures
5: autres (a specifier)

7.b- Appreciation de la qualité aprés semis :
1= Taux de germination
2= vigueur a la levée
3= Nombre de plant en bonne sante
4= Autres (a specifier)

8- Pre'férence variétale

8.a — N om de la variété dominante : écrire le nom

Temps d’acquisition : Nombre d’années

174

8.b- Description variétale :
Morphologie :
Taille : l= taille haute, 2= taille moyenne, 3= taille courte
Couleur desfeuilles : écrire comme sur le fiche d’enquéte
Forme de branches : méme chose que pour la couleur des feuilles
Cycle végétatif : Nombre de mois
Rendement : nombre de kg/ha si mentionne autrement prend l’information figurant sur la
fiche
Couleur des graines : se conforme a la fiche
Adaptation a la zone : 1= Bonne, 2= Moyenne, 3= Mauvais
Autres (a specifier)
8.d — Avez-vous change votre variete dominante : l= oui, 2= Nom
Si oui pourquoi : l= chute de rendement

2= irregularite des pluies

3= probleme culinaire

4= pauvrete des sols

5= autres (a specifier)

8.e- le caractéristique varietal qui vous attire plus pour cette variété
l= Rendement en graines 5= tolerance de la Sécheresse

2: Resistance aux insectes 6= utilisation secondaire

3= préoocité 7= Autres (a specifier)

4= aptitude culinaire
9- Méme codes que pour la question 3

Information générale :
Taille du champ = nombre d’hectare
Personnes travaillant dans ce champ plus de vous-meme: Nombre de personnes
Alle' a l’école : 1= oui 2: Non
Si oui, niveau atteint :

1= premier cycle

2= Second cycle

3: lycee/ecole secondaire

4: University

5= Alphabetisation en langue nationale

6= Autres (a specifier)

175

 

APPENDIX 11: Survey consent statement

Informed consent form for Michigan State University/ICRISAT study of seed
quality assessment for Sorghum Groundnut and Millet in two different agro
ecological Zone: Mande and Dioila in Mali Republic

This survey is part of a team effort of Michigan State University and ICRISAT national
of Mali in collaboration with: Association Conseil pour le Developpement (ACOD),
Association des Organization professionnelles Paysanne (AOPP) and the Office de la
Haute Vallee du Niger (OHVN) in Mande, and Union Local des producteurs de Cereales
(ULPC) in Dioila. It aims at assessing seed quality and farmers preference for varietal
traits through

0 Documenting seed quality through assessing samples from different sources.

0 Assessing farmers’ preference for varietal traits

0 Developing recommendations to improve farmers’ access to quality traits

The interview will take about one hour to complete and will be repeated. The first run
will focus on seed sources and collection of seed samples and the second will be on the
farmers’ choice for varieties based on varietal traits

If you choose to participate, you may refuse to answer any questions, or you may stop
participating at any time.

Your responses will be kept confidential to the maximum extent allowable by law. Your
responSes will be summed together with those of 270 other households in Mande and
Dioila and only results from analysis will be reported.

You indicate your voluntary consent by participating in this interview: May we begin?

If you have any questions or concerns for ICRISAT and Michigan State University
regarding your rights as a study participant, or are dissatisfied at any time with any aspect
of this study, you may contact - anonymously, if you wish, Rattunate Eva Weltzen, Mrs
Goita Marthe Diallo ICRISAT at (tél. 222 33 75), Abdoulaye Sangare for ACOD,
Mamadou Coulibaly for AOPP et Ibrahima Camara for OHVN in Mande and Mamoutou
Diarra for ULPC in Dioila and Peter Vasilenko, Ph.D., Director of the Human Research
Protection Programs (HRPP) at Michigan State University: (517) 355-2180, fax: (517)
432-4503, email: irb@msu.edu, or regular mail: 202 Olds Hall, East Lansing, MI 48824
USA

176

REFERENCES

ABT 2002. Mali Agricultural Sector Assessment. Volume 1: Summary of Finding and
Recommendations, Agricultural Policy Development Project. Abt Associates Inc.,
BetheSEa.

Almekinders C.J.M., NP. Louwaars and CH. Bruijn. 1994. Local seed systems and their
importance for an improved seed supply in developing countries. Euphytica, vol 78 #3,
pp 207- 216, June 1994 in International Journal of Plant Breeding.

Almekinders C.J.M. and NP. Louwaars. 1999 : Farmers’ seed production: New
approaches and practices Intermediate Technology Publication, London

Almekinders C.J.M. and W. de Boef. 1999 . The challenge of collaboration in the
management of crop genetic diversity. ILEIA Newsletter December 1999, pp. 5-7

Almekinders C.J.M. 2001. Management of Crop Genetic Diversity at Community level.
Deutshe Gesellschafi fur Technische Zusammenarbeit (GTZ) project of Managing Agro-
biodiversity in Rural Areas, Department of Rural Development Eschborn, Germany.

Almekinders C.J.M. & J aap Hardon (eds). 2007 : ‘Bringing Farmers Back Into Breeding :
Experiences with Participatory Plant Breeding and Challenges for Institutionalization.
International Development Research Center (lDRC) Ottawa

Baidu—Forson, J. 1997. On-station farmer participatory varietal evaluation: a strategy for
client-oriented breeding. Experimental Agriculture 33: 43-50

Basra . AS, 1995 : Seed Quality: Basic Mechanisms and Agricultural Implications.
Haworth Press.

Bazile, D. (2006). State-farmer partnerships for seed diversity in Mali. London:
Gatekeeper series 127.

Bellon, MR. 1996. The dynamics of crop infraspecific diversity: A conceptual
framework at the farmer level. Economic Botany 50:26-39.

Benech-Amold, Roberto L.; R. A. Sanchez. 2004 : Handbook of Seed Physiology
Application to Agriculture. Food Products Press Imprints of The Haworth Reference
Press, Inc New York. London. Oxford

Becker, LC. (2000). Garden money buys grain: Food procurement patterns in a Malian
village. Human Ecology: 28(2): 219-250.

177

Brick, MA. 2004. “Improve Yield with High Quality Seed.” Colorado State University
Cooperative Extension professor, soil and crop sciences. (publication details)

Broek. E. van den. 2007. Sorghum production by Malian women —

its role/importance. Results of a field research for MSc thesis in International
Development Studies; specialization Rural Development Sociology Wageningen
University

Boughton. D., B. H. De Frahan. 1994. Agricultural Research Impact Assessment the case
of Maize Technology Adoption in Southern Mali. MSU International Development
Working Paper No. 41. Department of Agricultural Economics and the Department of
Economics, Michigan State University, East Lansing.

Byth D.E. (ed.) , 1993. “Sorghum and Millets Commodity and Research Environments.”
Patancheru, AP. 502 324, India: lntemational Crops Research Institute for the Semi-Arid
(ICRISAT).

Community Technology Development Trust (CTDT) 2006. Community Seed Fairs in
Zimbabwe, presented in Community Based Development Center (CBDC) (40—43)

Chakanda. R, 2000. Farmers’ seed systems for sorghum in Mali: An evaluation of
farmers’ variety characterization criteria. MS Thesis, Wageningen University and
Research Centre, research conducted at ICRISAT Regional Centre at Bamako, Mali.

Chavas . J .P, P.M. Kristjanson and P. Matlon, 1991. “On the role of information in
decision making: the case of sorghum yield in Burkina Faso”. Journal of Development
Economics, Vol. 35, pp 261-280., Elsevier Science Pubishers B.V. (North-Holland)

Christinck. A, E. Weltzien, and V. Hoffmann (eds). 2005. “Setting Breeding Objectives
and Developing Seed Systems with Farmers”. A Handbook for Pratrical Use in
Participatory Plant Breeding Projects. Margraf Publishers, Weikershein, Germany and
CTA, Wageningen, The Netherlands

Cooper.H.D. 1 and L. Sperling 2. 2003. Understanding seed systems and strengthening
seed security. 1 Seed and Genetic Resource Service Food and Agriculture Organization
(FAQ), 2 lntemational Center for Tropical Agriculture (C IAT). Effective and Sustainable
Seed Relief: A Stakeholders’ Workshop, Rome, 26-28 May 2003

Coulibaly . S. 2004. Characteristics of the seed sector in the Republic of Mali. West
Africa Seed and Planting Material, the Newsletter of west Afiica seed Network
(WASNET) # 12 ISSN 1595-2312, PO Box 9698, KIA Accra, Ghana.

Copeland. L. O, M. B. McDonald. 2001: Principles of Seed Science and Technology.
Springer,

178

Daniel. I. O, J .A. Adetumbi. 2006. Maize Seed Supply Systems and Implications for Seed
Sector Development in Southwestern Nigeria. Journal of Sustainable Agriculture, Vol.28
(2)

Defoer, T., A. Kamara and H. De Groote. 1997. Gender and variety selection: Farmers’
assessment of local maize varieties in Southern Mali. African Crop Science Journal 5:65-
76

Diakite', S. 2004. Etude sur la diffusion des variétés ame'liorées de sorgho a partir de la
vente de semences dans la zone de Dioila. ICRISAT, Mali.

Diakité, S. 2003. Le systeme semencier local: Description, Evaluation, Valorisation. Cas
de 7 villages au Mali. Working papers on Local Knowledge #2. Point Sud, Bamako,
Mali.

Diakité . L and A. M. Diarra . 2000. Etude du développement de la filiere des semences
dans les pays du Sahel (cas du Mali). ClLSS/INSAH/IER. Bamako, Mali

Doss.C.R. 2001. Designing Agricultural Technology for African Women Farmers:
Lessons from 25 Years of Experience. Yale University, New Haven, CT, USA and
CIMMYT (International Center for the Improvement of Wheat and Maize) Mexico city,
Mexico.

Downess, R. W. 1970. Effect of light intensity and leaf temperature on photosynthesis
and transpiration in wheat and sorghum. Aust. J. Biol. Sci., 24:843

Folkertsma. R.T., H. F. W. Rattunde, S. Chandra, G. Soma Raju and C. Tom
Hash. 2005. The pattern of genetic diversity of Guinea-race Sorghum bicolor (L.)
Moench landraces as revealed with SSR markers. Theor Appl Genet (2005)

Elang. J, 2004. Commercial Channels of seed and Planting Materials in sub-Saharan
Africa: Seed markets and regulations in sub—Saharan Africa. West Afi‘ica Seed and
Planting Material, the Newsletter of west Africa seed Network (WASNET) # 12 ISSN
1595-2312, PO Box 9698,KIA Accra, Ghana

Food and Agriculture Organization of the United Nations (F.A.O.) .1997. “ Production
des semences de qualité déclarée au Rwanda. ” Kigali, Rwanda.

Food and Agriculture Organization of the United Nations (F.A.O.) (2004) : Seed fairs
revive a battery of a tradition.
http://www.tamargnewsroomz’en/focus/2004/40023.v/articlc 4‘)()3(’)en.html.

FAOSAT production/ProdStat/Crops,
http://lhostat.f'aonrg/sitc/567/DcsktopDcfauItaspx‘.’Pagcl DT—r5(r7#ancor

 

Fehr, W.R. 1987. Principles of cultivar development: Volume 1, Theory and technique.
Macmillan Publishing Company. New York.

179

Friis-Hansen .E, and Turner .M, 1992. The seed sector in developing countries: A
framework for performance analysis. Overseas Development Institute (ODI). working
paper #65, Regent’s college Inner Circle, Regent’s Park London NW1 4NS

Harlan JR. Agricultural origins: centers and noncenters. Science (1971) 174:468—474

Heisey, P.W. and J .P. Brennan (1991) An Analytical Model of Farmers’ Demand for
Replacement Seed, American Journal of Agricultural Economics. 73(4): 1044-52

Huvio.T and A. Sidibe. 2003 Strengthening farmers’ capacities for plant genetic
resources conservation in Mali. Plant Genetic Resources l(l ); 31-41

ICRISAT, October 2005. Creating opportunities for poor farmers: Intensification of
Sorghum and Millet Systems using local biodiversity and market opportunities in Semi-
Arid West Africa. A research proposal for Sahelian and Sudanian zones of Mali, Niger,
Burkina Faso and Nigeria, Bamako Mali.

ICRISAT 2002. Project Proposal, ICRISAT, Mali

J affee, S.and J. Srivastava 1992. Seed Development: The appropriate Role of the private
and pubic sectors. World Bank Discussion paper. # 167. Washington, DC, World Bank.

Jarvis, D. I., A. H. D. Brown, P. Hung Cuong, L. Collado-Panduro, L. Latoumerie-
Moreno, S. Gyawali, T. Tanto et al.. 2008. A global perspective of the richness and
evenness of traditional crop-variety diversity maintained by farming communities.
Proceedings of the National Academy of Sciences, USA 105: 5326—5331

Jarvis DI, and Campilan DM (2006) Crop Genetic Diversity to Reduce Pests and
Diseases On-F arm: Participatory Diagnosis Guidelines (Bioversity lntemational, Rome)
Ver 1.

Jayne .T.S. Isaac J. and Gem Argwings-Kodhek, 2002. Perspectives on Agricultural
Transformation: A view from Afi‘ica. Nova Science Publishers, Inc New York.

Kelly, V.; Carpenter, J .; Diall, O.; Easterling, T.; Koné, M.; McCormick, P.; McGahuey,
M.; 2005: Options for Economic Growth in Mali through the Application of Science and
Technology to Agriculture. USAID.

Kudadjie. C.Y; P.C. Struik; P. Richards and SK. Offei. 2004. Assessing Production
Constraints, Management and use of Sorghum diversity in North-east Ghana: A

diagnostic study. NJAS # 52-3/4 PO. Box 68 Legon, Ghana.

Leonardo. W. J. 2002. Organizing seed fairs in emergency situations: Improving the
efficiency of seed distribution. lntemational Crops Research Institute for the Semi-Arid

180

Tropics (ICRISAT) and Instituto Nacional de Investigacao Agronomica (INlA) Maputo
Mozambique.

Longley .C. 2004. Relief or Agricultural Development? Emergency seed projects,
farmers seed systems and the dissemination of modern varieties in Mali and Niger.
Working Paper. lntemational Crops Research Institute for the Semi-Arid Tropics
(ICRISAT) Patancheru 502 324, Andhra Pradesh, India.

Maredia . M, Howard .J, and Boughton .D, with Naseem .A, Wanzala .M, and Kajisa .K
.1999. “Increasing Seed System Efficiency in Africa: Concepts, Strategies, and Issues”.
MSU lntemational Development Working Paper #77. Department of Agricultural
Economics, Michigan State University, East Lansing, MI

Manderscheid, L. 1965. Significance Levels-005, 0.01, Journal of Farm Economics
47:1381-1385.

Mburu . C. 1986. Importance of Genetic Resources and Breeding of Small Millets in
Kenya, In Proceedings of the First lntemational Small Millets Workshop, Bangalere,
India, October 29 to November 2 1986. small millets in global Agriculture, pp. 149-154.

Mekbib. F. 2006. Farmer and formal breeding of sorghum (sorghum bicolor (L) Moench)
and the implications for integrated plant breeding. Euphyca (2006) 152:163-176

Minot . N, M. Smale, C. Eicher, T. Jayne, and J. Kling,. 2006. Seed development
programs in sub-saharian Afi'ica: A review of experiences. Submitted to Peter Malton
Rockefeller Foundation Nairobi, Kenya 11 November 2006.

Mulatu. E. and K. Belete. 2000. Participatory Varietal selection in Lowland Sorghum in
Easthem Ethiopia: Impact on Adoption and Genetic Diversity. Alemaya University of
Agriculture (AUA), Department of Plant Sciencs PO. Box 138, Dire Dawa, Ethiopia.
Expl Agric (2001), Vol 37, pp. 211-229 printed in Great Britain.

Nigam S. N., Giri D.Y. and Reddy A.G.S. 2004. Groundnut Seed production Manual.
Patancheru 502 324, Andhra Pradesh, India: ICRISAT. 32 pp.

Nkongolo. K.K., K. K. , L Chinthu, M. Malusi, and Z. Vokhiwa. 2008. Participatory
Variety Selection and Characterization of Sorghum (sorghum bicolor (L) Moench) elite
accessions from Malawian genepool using farmer and breeder knowledge. Afiican
Journal of Agricultural Research Vol. 3(4), pp. 273-283 ISSN 1991-637X 2008
Academic Journal

N’tare, B.R.; A.T. Diallo.; J. Ndjeunga and F. Waliyar. 2008 . Groundnut Seed
Production Manual. Patancheru 502 324, Andhra Pradesh, India: lntemational Crop
Research Institute for Semi-Arid Tropics (ICRISAT)

181

N’tare, B.R.; A.T. Diallo.; J. Ndjeunga; F. Waliyar; O. Kodia; CA Echekwu; I. Kapran;
A.D. Sylva; A.T. Diallo; A. Amadou; H. Y Bissala and K. Sako . 2008 : Farmer
Participatory Evaluation and Dissemination of Improved Groundnut Varieties in West
Africa. lntemational Crop Research Institute for Semi-Arid Tropics (ICRISAT)
publication Andhra Pradesh, India. 30pp

Omanya. G.O., E. Weltzien-Rattunde, D. Sogodogo., M. Sanogo, N. Hassens, Y. Guero
and R. Zangre. 2006. Participatory Varietal Selection with Improved Pearl Millet in West
Afi‘ica. Experimental Agriculture (2007), volume 43, pp. 5-19 2006 Cambridge
University Press

Puchner. L. 2003. Women and literacy in rural Mali: a study of socio-economic impact of
participating in literacy programs in four villages. lntemational Journal of Education
Development 23 (2003) 439- 458. Elsevier Science Ltd

Rao, S.A., K.E. PrasadaRao, M.H. Mengesha&V.Gospal Reddy, 1996. Morphological
diversity in sorghum germplasm from India. Genetic resources and Crop Evaluation
43:559-567.

Rasabandit. S., P. Jaisil, G. Altin, C. Vera Cruz, B. Jongdee, and P. Bantemg. 2006.
Participatory Varietal Selection (PVS) to Assess Farmer Preferences of Traditional
Glutinous Rice Varieties in the Lao PDR. KHON KAEN AGRIC. 34 (3): 128-141 (2006)

Rice, E. M. Smale and J. L. Blanco. Farmers use of improved seed selection practices in
Mexican maize: evidence and issues from the Sierra de Santa Marta. World Development
26:1625-1640

Saito. K. A and C. Jean Weidemann. 1990 Agricultural Extention for Women Farmers in
Afiica. Population and Human Resources Department, The World Bank Working papers
# 0398 Women in Development.

Saito. K. A., H. Mekonnen, D. Spurling, 1994. Raising the Productivity of Women
Farmers in Sub-Saharan Africa

Shargie .N.G and W.Wenzel. 2008 . Identifying F armers’ preferences for improved
sorghum varieties in Limpopo, South Africa. ARC-Grain Crop Institute, Potchefstroom,
NW, 2520, South Afi‘ica.

http://www.sik.se/traditional grains/review/ PosterO/620Presentati01159-1320PDF‘IQZOfiIes/Sha

rgiepdf

 

SAS GLIMMIX PROCEDURE documentation. June 2006. online resource
htgfl/suppurt.sas.com/md/app,/pap_ers/glimmixpdf

 

SAS GLM CONTRAST PROCEDURE. 1999 SAS OnlineDoc®, Version 8, Cary, NC:
SAS Institute Inc. North Carolina 27513

182

http://www.sfu.ca/saS Eoc/sashtml/ stat/CImp3I )/ sect I 2.htm

 

Setimela. P. S., Emmanuel Monyo. 2004: Successful Community-based Seed Production
Strategies. lntemational Maize and Wheat Improvement Center, Marianne Béinziger
Contributor lntemational Maize and Wheat Improvement Center . CIMMYT.

Shiferaw. B.A, T.A.Kebede, L. You. 2008. Technology adoption under seed access
constraints and the economic impacts of improved pi geonpea varieties in Tanzania.
Agricultural Economics 39 (2008) 1-15.

Siart, Sonja. 2008 : Strengthening Local Seed Systems: Options for Enhancing Diffusion
of Varietal Diversity of Sorghum in Southern Mali. Margraf Publishers, Scientific books.
KanalstraBe 21; D-97990 Weikersheim .

Siart, S., E. Weltzien, M. Kanoute, and V. Hoffmann. 2006. Farmers’ Sorghum Seed
Sources afier a Drought in Southern Mali. Conference Proceedings Tropentag
Conference on lntemational Agricultural Research for Development, University of Bonn,
Germany.

Scott. J, P. Kambewa, R. Chirwa, and V. Aggarwal revised by Soniia David. 2003 : Local
seed systems for beans in Malawi CIAT Africa Occasional Publication series # 40
Kampala, Uganda.

Sperling .L, E. Weltzien, M.B. Sangare, J .S. Shines, S.S. Bore, A. Bamba, C.Traore, D.
Coulibaly, S. Traore, C.O. Keita, M.A. Hamada, M. Ballo, F. Sangare, M. Kanoute, B.
Sanogo, H. Guindo, B. Konte, S. Sanogo, A. Traore, M. Leoffen and A. Dembele . 2006.
Seed system security assessment (SSSA) in DOUENTZA, Northern Mali. Catholic Relief
Services/Mali and Partners, Bamako Mali. Internal Report

Tefft, J. 2004: Mali’s white revolution: Smallholder cotton from 1960 to 2003. Focus 12,
Brief 5 of 10, Building on Successes in African Agriculture 2020 Vision for Food ,
Agriculture and Environment, IF PRI, Washington.

Tripp .R, and Rohrbach .D. 2001.Policies for African seed enterprise development.
Overseas Development Institute (ODI), 111 Westminster Bridge Road, London SE1 7J D,
UK. lntemational Crop Research Institute for Semi-Arid Tropics (ICRISAT), P.O.776,
Bulawayo, Zimbabwe. Food Policy 26 (2001) 147-161.

Tripp. R. 2000. The organization of farmer seed systems. Relevance for participatory
plant breeding. Overseas Development Institute (ODI), paper prepared for the
symposium, “Scientific Basis for Participatory Improvement and Conservation of Crop
Genetic Resources”, Oaxtepec, Mexico, 8-14 October 2000.

Tripp, R. (1997) The Institutional Conditions for Seed Enterprise Development. Overseas
Development Institute (ODI), Working paper 105, London

183

Venkatesan .V, 1994. Seed Systems in Sub-Saharan Afiica: Issues and Options. World
Bank Discussion papers Africa Technical Department series # 266, Washington, DC.
World Bank.

Vemooy.R. 2003. Seed that give: Participatory Plant Breeding. lntemational
Development Research Center (IDRC) PO Box 8500, Ottawa, ON, Canada KlG 3H9
ISBN 1-55250-014-4

Weltzien, E.; Christink, A.; Touré, A.; Rattunde, H.F.W.; Diarra,M.; Sangare, A.;
Coulibaly, M. 2006. : Enhancing farmers’ access to sorghum varieties through scaling-up
participatory plant breeding in Mali, West Afiica. In Almekinders, C.; Hardon, J. (Eds)
2006: Bringing Farmers Back into Breeding. Experiences with participatory Plant
Breeding and Challenges for Institutionalization. Agromisa Special 5, Agromisa,
Wageningen.

E. Weltzien, K. Von Brocke. 2001. Seed systems and their potential innovation:
Conceptual framework for analysis. In Sperling. L. 2001. Targeted Seed Aid and seed
system interventions: Strenghtening small farmer seed systems in East and Central
Afiica. Proceedings of Workshop held in Kampala, Uganda, 21-24 June 2000.

Witcombe. J .R., A. J ashi, K.D. J ashi and B.R. Sthapit. (1996). Farmer Participatory Crop
Improvement 1. Variety Selection and Breeding methods and their impact on biodiversity.
Experimental Agriculture 32: 445-460.

Wolfgang Pejuén. 2005. Constraints to and opportunities for developing a bean seed
production and marketing system in Honduras. Thesis submitted to Michigan State
University in partial fulfillment of the requirements for the degree of Master of Science in
the Department of Agricultural Economics.

Wolfinger, R. and O’Connell, M. (1993), “Generalized Linear Mixed Models:
A Pseudo-Likelihood Approach”, Journal of Statistical Computation and
Simulation, 4, 233—243.

Yapi,A.M.; Kergna,A.O.; Debrah,S.K.; Sidibe,A.; Sanogo,O. 2000. Analysis of the
economic impact of sorghum and millet research in Mali. Impact-Series-International-
Crops-Research-Institute-for-the-Semi- Arid-Tropics (India).no. 8, 57 p. ICRISAT—
WASIP, BP 320, Bamako, Mali.

Web link:

http://www.ciagow’bizv’publ ications/factbtmk«’gcos/mI.lrtml

 

http://tlata.un.orgv’CmmtrvafiIe.aspx‘?crNarrre:Mali

 

http://cn.wikipcdiabrg/wikis’POl I i nation

 

184

http://www. icrisat.or‘g’sm‘glrurn/sorghum.htm

 

185

   

”ll11111111111311“[Illllliljiljljlljllles