World Development Vol. 35, No. 9, pp.

1579–1593, 2007
Ó 2007 Elsevier Ltd. All rights reserved
0305-750X/$ - see front matter
www.elsevier.com/locate/worlddev
doi:10.1016/j.worlddev.2006.05.023

The Dynamics of Farmers’ Maize Seed

Supply Practices in the Central Valleys of
Oaxaca, Mexico
LONE B. BADSTUE
International Maize and Wheat Improvement Center (CIMMYT), Mexico

MAURICIO R. BELLON
International Maize and Wheat Improvement Center (CIMMYT), Mexico
Bioversity International, Maccarese, Italy

JULIEN BERTHAUD
International Maize and Wheat Improvement Center (CIMMYT), Mexico
Institut de Recherche pour le Développement (IRD), Montpellier, France

ALEJANDRO RAMÍREZ, DAGOBERTO FLORES

International Maize and Wheat Improvement Center (CIMMYT), Mexico

and
XÓCHITL JUÁREZ *
Universidad Autónoma de Chapingo (UACH), Texcoco, Mexico
Summary. — What are the organizing principles that underlie and help shape farmers’ seed supply
practices in the Central Valleys of Oaxaca? And what are the implications of these practices for
maize genetic diversity and the introduction of improved varieties? Local maize seed supply was
studied using both qualitative and quantitative methods and a series of factors that influence local
seed supply was analyzed. Together they constitute a set of flexible and dynamic practices, which
embrace both conservation and innovation aspects. Implications for on-farm conservation and
introduction of improved varieties are discussed.
Ó 2007 Elsevier Ltd. All rights reserved.

Key words — Latin America, Mexico, maize, social seed system, on-farm conservation

* This paper builds on research carried out with funds from the Systemwide Program on Collective Action and
Property Rights (CAPRi) of the Consultative Group on International Agricultural Research (CGIAR), as well as
with funds from the government of France, the government of Denmark, and the Food and Agriculture Organization
of the United Nations (FAO). The authors thank Kevin Pixley, Jens Riis-Jacobsen, and three anonymous reviewers
for their comments on earlier drafts of this paper, as well as the farmers who participated in this research for the
information they provided and for their kindness and patience. We also wish to thank Jillian Baker for much-
appreciated editorial assistance. Final revision accepted: May 6, 2006.
1579
1580 WORLD DEVELOPMENT

1. INTRODUCTION local crop genetic resources and make special

use of diverse crop varieties in their production
Seed is a fundamental input for agriculture system. Decisions regarding varietal choice
and probably the single most important input often depend on multiple considerations and
in all crop-based farming systems. More than not just yield. Such issues can be illuminated
being the basis of production for the majority by in-depth, qualitative studies (Orlove &
of the world’s crops, seed determines the upper Brush, 1996).
limit on yield and therefore on the ultimate pro- Mexico is a center of domestication and
ductivity of all other inputs (Cromwell, 1990). diversity for maize (Matsuoka et al., 2002;
Securing access to seed of the desired varieties Piperno & Flannery, 2001; Sanchez, Good-
and of good quality is therefore a very impor- man, & Stuber, 2000), and farmers continue
tant issue for farmers and a concern for society to play a key role in the maintenance of this
to achieve food security. diversity (Bellon, 2004; Hernandez, 1985;
Informal seed systems 1 are still the prevail- Perales-Rivera, Brush, & Qualset, 2003).
ing source of seed in developing countries, According to Morris and López-Pereira (1999)
and many studies have stressed their impor- approximately 80% of the area planted to maize
tance (Almekinders, Louwaars, & de Bruijn, in Mexico is planted with recycled seed, that is,
1994; Seboka & Deressa, 2000; Sperling, seed selected by farmers from the previous har-
Heidegger, & Buruchara, 1995; Subedi, Chau- vest.
dhary, & Sthapit, 2003; Thiele, 1999). Several The structure and evolution of maize genetic
authors have pointed out that informal systems diversity in the Central Valleys of Oaxaca de-
are mostly based on traditional social alliances pend on a combination of gene flow and farm-
and family relations, cast in the context of mu- ers’ selection (Pressoir & Berthaud, 2004).
tual interdependence and trust, often forming While gene flow allows new genes to enter the
dynamic networks with a high degree of com- system, farmers’ seed selection, which is
plexity (Almekinders et al., 1994; Seboka & strongly influenced by local preferences and
Deressa, 2000). Nevertheless, relatively little is culture, allows for the differentiation between
known about how these systems function. As varieties from the same farmer or between
stated by Seboka and Deressa (2000, p. 250): farmers (Pressoir & Berthaud, 2004). 2
‘‘The flow of seeds or farmer-to-farmer ex- The dynamics of farmers’ seed supply prac-
change of seed is a neglected area of research. tices have important implications, both for
There is an urgent need to understand more the conservation of crop genetic resources on-
in detail the process of farmer-to-farmer ex- farm and for the design and implementation
change of seed.’’ of interventions to support conservation (Bel-
Seed is also an important source of germ- lon, 2004; Subedi et al., 2003). Moreover, in a
plasm for crop improvement, and a key element broader perspective they also have important
in the management and conservation of crop implications for the introduction of new varie-
genetic resources (Almekinders, 2001; Almekin- ties and seed sector development.
ders & de Boef, 2000; Bellon, 2004; Jarvis et al., This paper provides an overview of a range of
2000; Orlove & Brush, 1996). Access to a wide factors influencing farmers’ local seed transac-
range of genetic diversity allows farmers and tion practices. 3 Together, these constitute a dy-
plant breeders to adapt a crop to heterogeneous namic set of practices which, on the one hand,
and changing environments, developing culti- ensure an efficient and low-cost supply of a di-
vars with high levels of adaptation to biotic verse array of maize germplasm to farmers in
and abiotic stresses and to human preferences. the study communities, and, on the other hand,
Consequently conservation of genetic resources are efficient in maintaining and conserving
is crucial to address future challenges in re- maize genetic diversity at the local level.
search and crop improvement in order to inten-
sify agricultural production and increase food
supply, and to try to respond to farmers’ differ- 2. STUDY SITES AND METHODOLOGY
ent requirements and preferences.
Studies on these aspects often portray farm- In the Central Valleys of Oaxaca, maize agri-
ers as the caretakers of important crop genetic culture continues to play a significant role as a
resources. However, most farmers do not main- source of food security, local maize products,
tain crop genetic resources merely for the sake and income. Farmers in this region value their
of conservation. Rather, many farmers value landraces and continue to plant them, thereby
 THE DYNAMICS OF FARMERS’ MAIZE SEED SUPPLY PRACTICES 1581

contributing to the conservation of maize bio- of irrigated maize, was the area planted to im-
diversity (Bellon et al., 2003; Smale et al., proved maize seed, slightly significant (Smale
2003). A formal seed sector has yet to develop et al., 1999).
in this region, and most farmers produce their The population in the study area is predomi-
own maize seed year after year or rely on other nantly Spanish speaking, but in Santa Ana Zeg-
farmers to acquire seed. ache and Santo Tomás Mazaltepec more than
The study presented here was carried out in 30% belong to the Zapotec ethnic group and
six communities in the Central Valleys of Oax- speak Zapotec as a first language. In both com-
aca: San Pablo Huitzo, Santo Tomás Mazalte- munities more than 97% of Zapotec speakers
pec, San Lorenzo Albarradas, San Agustı́n also speak Spanish (INEGI, 2001a).
Amatengo, Valdeflores, and Santa Ana Zeg- Agriculture is the major source of income for
ache (Table 1). Average farm size is 3.5 ha, nearly all households, whether located in better
and farming systems in all six communities or poorer maize production zones. Nonfarm
are characterized by low productivity (Smale employment is an important source of income
et al., 1999). Yearly mean temperature in the re- for about one-quarter of the farmers. A similar
gion is 18–22 °C, with an average annual pre- overall percentage of households depend on
cipitation of 600–1000 mm (INEGI, 2001b). remittances, although the percentage varies
Maize, beans, and squash are the most com- considerably among some of the study commu-
mon crops, with maize being sown on average nities, particularly in Huitzo, Mazaltepec, and
on more than 90% of the landholding areas. Amatengo (Smale et al., 1999).
Maize agriculture is dominated by local land- As the CIMMYT/INIFAP project was con-
races: only in the area of San Pablo Huitzo, ducted in the same area, considerable back-
which also had the highest average percentage ground information on the study communities

Table 1. Key characteristics of the six study communities

Characteristics Communities
San Pablo Santo San Lorenzo San Agustin Valdeflores Santa Ana
Huitzoa Tomas Albarradasa Amatengo Zegachea
Mazaltepec
Populationb 4,685 1,939 1,857 1,752 1,246 2,543
Altitude (masl)c 1,700 1,660 1,810 1,360 1,460 1,480
Maize yield potential Good Poor Poor Poor Good Good
Mean
No. farmer varietiesd/ 1.26 1.21 2.13* 1.10 1.11 1.98*
household in 1997
Farm size 1996 (ha) 2.44** 3.91 4.01 2.84** 3.87 3.46
% Land privately owned 49.60* 0.00 1.00 27.42* 0.00 100*
% Maize area irrigated 54.2* 15.7 8.10 11.90 3.78** 0.17**
% Maize area in improved seed 0.14* 0.00 0.04 0.00 0.00 0.01
Percent
Households dependent 40* 28 30 25 25 15*
on local nonfarm
income
Households dependent 3* 13* 23* 38 25 25
on remittances
Sources: Smale et al. (1999) and INEGI (2001a).
a
Communities where the seed flow tracer study took place.
b
Total population in relevant localities within the municipality (INEGI, 2001a).
c
Altitude of municipal center (INEGI, 2001a).
d
Crop populations that a group of farmers recognize as distinct units. A farmer variety is not a variety in the sense of
commercial agriculture, where a variety should be distinct, uniform and stable.
*
Mean (frequency) significantly higher (different) using one-tailed t-test (chi-squared test), .05 significance level.
**
Mean significantly lower using one-tailed t-test .05 significance level.
1582 WORLD DEVELOPMENT

was already available, including a random sam- transactions. Focus group results were remark-
ple of 240 households. The communities were ably similar across communities and gender.
selected for the contrasts they represented in For the survey-based seed flow tracer study,
terms of maize yield potential and dependency which involved male and female representatives
on nonfarm income (Smale et al., 1999). For from 153 farm households, we focused again on
all the activities of this study, informants were the same three communities where the initial
selected based on information from the previ- interviews were conducted. This was justified
ous study representing different age groups, because the results of the focus group discus-
gender, ethnicity, economic status, and level sions suggested that the conditions of these
of formal education to ensure that the diversity three communities were representative for all
of different social groups was captured. six, and because the tracer study was very labor
The original objective was to determine the intensive and the resources limited. In the tra-
role of collective action in on-farm conserva- cer study, we followed the flows of seed among
tion of maize genetic diversity. However, as re- selected farm households, paying special atten-
search progressed no evidence of institutions of tion to farmers’ explanations about the transac-
collective action specific to seed acquisition was tions: why they had engaged in a transaction,
identified. As a consequence the research focus with whom, and what the significance of the
was redirected toward a more general analysis transaction was, among other factors. A total
of how farmers access seed of diverse maize of 516 transactions of both incoming and out-
varieties, and the reasons behind the apparent going flows were recorded. Using criteria simi-
lack of collective action in this particular re- lar to those described for the selection of
spect (Badstue et al., 2006). informants, 10 households in each community
The research employed both qualitative and were selected as the point of departure. After
quantitative methodologies, including in-depth, a first round of interviews, households that
semi-structured ethnographic interviews, gave or received seed from each of the original
including an initial survey with key informants 30 households were located and queried in a
to identify relevant issues and questions; focus second round of interviews, and so on, until
group discussions; and a quantitative tracer each of the original households had led to an
study of seed flows among farm households. average of four additional households being
The different methods complemented each visited. 4
other and allowed key issues to be addressed
from several angles.
Early in the research process, in-depth ethno- 3. SEED
graphic interviews were conducted with 22 key
informants from the three most contrasting of (a) The distinction between grain and seed
the six communities in terms of maize yield po-
tential, dependency on nonfarm income, and From a biological perspective, any healthy
ethnicity (Smale et al., 1999): Santa Ana Zeg- maize kernel could serve as either seed or grain
ache, San Lorenzo Albarradas, and San Pablo for consumption. However, farmers in the
Huitzo, as part of an initial assessment of local study area clearly distinguish kernels as seed
practices for accessing seed of diverse maize for planting or as grain for consumption or
materials. These first interviews were carried sale. ‘‘Seed’’ represents a portion of the kernels
out in these three communities only, due to from the farmers’ harvest that has been care-
the labor intensive and time consuming charac- fully selected based on a set of specific criteria,
ter of the methodology. The data gathered dur- according to which farmers decide from which
ing these interviews were later confirmed in the ears to select the kernels to be used as seed,
focus groups discussions and in the tracer as well as which specific kernels on these ears
study. to define as seed. As documented by Smale
Twelve focus group discussions were carried et al. (1999), farmers’ seed selection criteria
out—one with men and one with women in tend to emphasize aspects related to ear and
each of the six communities. In total, 46 women grain health and size, and grain filling. How-
and 58 men participated. These discussions, ever, other factors may also play a role, for
which provided a great amount of valuable example, grain color or other local perceptions
information, covered the relative importance associated with what makes a ‘‘good’’ seed.
of seed loss as a vulnerability factor faced by In this process farmers exercise selection in
farmers and the mechanisms that guide seed an attempt to enhance favored varietal traits
 THE DYNAMICS OF FARMERS’ MAIZE SEED SUPPLY PRACTICES 1583

and lessen the influence of undesired ones. This this until the seed is planted and the maize
ensures that certain traits are passed to the next develops.
generation at a higher frequency. Furthermore, Environmental factors play an important
these traits are what define a variety in the eyes role in crop performance, however, some crops
of farmers, and studies have argued that this respond stronger than others across different
selection process plays an important part in environments. Maize exhibits what plant breed-
what structures diversity in farmers’ fields (Bel- ers call a high genotype-by-environment inter-
lon & Brush, 1994; Louette, Charrier, & Bert- action, meaning that its performance across
haud, 1997; Pressoir & Berthaud, 2004). As different agro-ecological environments depends
demonstrated by Pressoir and Berthaud (2004) on its specific genetic make-up. In other words,
the dynamics of maize genetic diversity is a a genotype or maize ‘‘variety,’’ which performs
combination of gene flow and selection; with- well in one environment, may not do so in
out farmers’ seed selection, maize populations another. 5
in this region would not show the great mor- To some degree seed quality is also subject to
phological diversity observed. the issue of nontransparency. Seed quality is
Meanwhile, smallholders in the Central Val- constituted by a range of factors and can be dif-
leys of Oaxaca generally refer to maize kernels ficult to assess, in particular the seed’s ability to
as grano or grain, without specifying its in- germinate. Age, pathogens, or inappropriate
tended use. Grain can be used for human con- storage may affect germination, but these fac-
sumption, animal feed, or sale. It has not yet tors are not necessarily visible to the human
been classified according to its intended use. eye.
Its destiny is therefore not unequivocal. By con- Even though they inspect the seed before
trast, once seed has been selected it is destined acquisition, farmers depend largely on the qual-
specifically for planting, and hence treated in ity of the information offered by the seed pro-
a different manner. vider with regards to traits and consumption
In spite of the seemingly clear distinction be- characteristics, environmental adaptation, and
tween seed and grain, farmers may sometimes seed quality. Farmers in the study area are
use grain as seed. This occurs mainly in relation aware of both the lack of transparency of seed
to smaller quantities of grain or during circum- and the genotype-by-environment characteris-
stances when it is difficult to obtain seed, for tic. One clear indication of this was, for exam-
example, due to lack of resources. Under these ple, that during the focus group discussions,
circumstances a farmer may decide to acquire adaptation to local agro-ecological conditions
grain rather than seed, and subsequently select was one of the first things farmers in all study
seed from this. However, as grain is generally communities mentioned as important when
managed less rigorously than seed, this proce- acquiring maize seed (Badstue, 2004; Badstue,
dure can entail additional risks with regards Bellon, Juárez, Manuel Rosas, & Solano,
to seed quality. Although a clearly defined con- 2003).
cept of seed exists (selected, clean, and of good
quality), it is not rigid or static. Rather, the
concept of seed is dynamic and negotiable, 4. USING OWN SEED
depending on the circumstances. This demon-
strates the flexibility in farmers’ categories The foundation of maize seed supply in the
and inclination toward experimentation and study communities is farmers’ practice of select-
practical solutions. ing seed from the previous harvest and saving it
for the next planting season. Of the farmers
(b) The lack of transparency in seed who participated in the seed flow tracer study,
75.8% relied entirely on their own seed in
Seed of good quality and with adequate 2001. Furthermore, Smale et al. (1999) reported
production characteristics for the particular that approximately 90% of all seed lots in the
location is fundamental for agriculture; how- study communities were selected by farmers
ever, these aspects can be difficult to assess from the previous harvest, while the rest were
when acquiring seed. Seed is not transparent acquired almost entirely from other farmers.
(Morris, Rusike, & Smale, 1998)—the traits According to informants, both in focus
and performance of the plants that will grow group discussions and individual interviews,
from it cannot be assessed by merely looking selecting and saving seed provides a sense of
at the seed. In principle, one cannot know security, as well as a chance to save money.
1584 WORLD DEVELOPMENT

Once seed is selected and safely set aside, one pride of being a good farmer—it is like a humil-
can rest assured that the seed for the next plant- iation!’’ On the other hand, though, as became
ing season is secured. Furthermore, the seed clear during the focus group discussions and
will be available when it is needed so that the the individual interviews, it is acceptable and
farmer will not incur planting delays. One can legitimate to obtain seed from other farmers in
therefore avoid spending money and/or time a bad year or for experimentation with other
acquiring seed at the last moment before plant- kinds of maize germplasm, provided one is gen-
ing, which is when prices typically increase and erally thought to manage seed with appropriate
many households are struggling to raise the care. In this case, the seed receiver has a justifi-
means necessary for land preparations, plant- able need for the seed, and is not someone who
ing, etc. prefers to rely on others for seed rather than
Farmers’ seed selection practices in the study make the effort of selecting and storing seed
area reflect both the genotype-by-environment from the previous harvest. In other words, this
consideration and the issue of seed security: person ‘‘deserves’’ the seed and will appreciate
Knowing the performance of the plants the the favor.
seed came from; farmers select maize seed Clearly, for farmers in the study area, select-
according to a set of characteristics that they ing and saving seed is not just a question of sav-
perceive as favorable in terms of their own ing money, but a decision that has cultural,
particular needs. Due to social, cultural, and economic, and agro-ecological components
environmental conditions, a variety that is (Badstue et al., 2005).
appropriate for one farmer may not necessarily It should be noted that, although farmers se-
be appropriate for another. Hence, what better lect their own seed year after year, they may
option to fit one’s needs and preferences than also, occasionally, substitute entirely, comple-
using the seed that one knows and has selected? ment, or mix their own seed with seed from
Moreover, for some of these farmers, their external sources. Initially a farmer may state:
own maize seed is associated with a certain ‘‘I have planted this white [maize] for 20 years.’’
affection value (Badstue et al., 2005). This as- However, further conversation may well reveal
pect surfaced many times during individual that on one or more occasions the seed was
interviews, and was also mentioned by farmers complemented or mixed with external seed.
during focus group discussions. Seed is often With regards to the study communities, these
inherited, passed on from parents to children practices have also been noted by Smale et al.
when the latter start farming independently. (1999) and similar practices have been reported
Often, the seed has been in the family for many from other regions in Mexico (Aguirre Gómez,
years during which it has provided the suste- 1999; Louette et al., 1997). Over time, these and
nance of the family, whereby it has acquired other management practices, for example, how
an inherent affection or symbolic value. Thus, the farmer selects seed, as well as naturally
for many farmers in the Central Valleys, the occurring pollen flow 6 from other farmers’
maize seed lot is something they have in-trust, maize fields, may well change the genetic
which links them with previous generations, make-up of his/her maize.
and which they, in turn, must pass on to their
descendants.
Finally, saving seed is strongly associated 5. SEED EXCHANGE
with being ‘‘a good farmer’’. In their own way,
each of the above mentioned aspects is part of Although saving seed from one’s own harvest
what constitutes the local concept of ‘‘a good is the backbone of local seed supply in the
farmer’’ a notion which can be said to lay out study area, farmers do acquire seed from other
certain principles for what is considered appro- sources from time to time.
priate behavior of a good farmer (Badstue et al., The quantity of seed involved in farmer-to-
2003). This should not be understood in a fixed farmer seed transactions is often quite small
or prescriptive sense, but rather as a set of (Table 2). Seed quantities were recorded in
abstract guidelines open to individual interpre- 386 transactions in the tracer study. While the
tation and negotiation. One aspect of appropri- average quantity was 12.5 kg, half of these
ate behavior of a ‘‘good farmer’’ is to take good transactions involved only 8 kg or less. In the
care of his/her seed (Badstue et al., 2003, 2005). CIMMYT/INIFAP research project, 2,726 kg
As the female farmers in one of the focus groups of seed of diverse maize varieties were sold to
stated: ‘‘[losing seed]. . . is like hurting one’s a total of 371 farmers, and the average quantity
 THE DYNAMICS OF FARMERS’ MAIZE SEED SUPPLY PRACTICES 1585

Table 2. Quantity of seed involved in transactions Several authors have argued that actors’ purpo-
Seed quantity per No. of transactions % sive behavior is embedded in concrete contexts,
transaction kga including systems of social relations, and there-
fore should be analyzed as such (Gudeman,
64 100 25.9
2001; Long, 2001; Swedberg & Granovetter,
5–8 93 24.1
2001).
9–12 59 15.3
In a recent paper Granovetter points out
13–16 43 11.1
why social structure should be regarded as an
17–20 42 10.9
important influence on economic outcomes
21–40 45 11.7
41–48 4 1.0
(Granovetter, 2005). First of all, social net-
works affect the flow and the quality of infor-
Total 386 100 mation to a significant degree. At the same
a
Unknown for 130 of the 516 transactions. time they constitute an important source of re-
ward and punishment, which often has a bigger
impact when coming from others personally
purchased was 4.3 kg (Bellon, 2004). In com- known and whose acceptance we seek. Finally,
parison, farmers in the region normally calcu- Granovetter notes, that where trust emerges, it
late four almudes 7 of seed, approximately does so in the context of a social network.
16 kg, to plant 1 ha maize. In spite of the fact, The presence of trust can provide a more se-
that most plots are very small in the Central cure environment for transactions and social
Valleys and therefore seldom require large exchange, as demonstrated, for example, by
amounts of seed, the high percentage of seed DiMaggio and Louch (1998) in a study of peo-
transactions involving small quantities of ple’s use of networks in relation to a certain
seed suggests that a considerable part of all range of consumer transactions. The authors
seed flows are motivated by elements of farmer conclude that uncertainty about product char-
experimentation, or take place to complete the acteristics or performance quality leads people
required amount of seed in the event of partial to prefer sellers with whom they have noncom-
seed loss. mercial ties. This is effective, they argue, be-
It is difficult to assess the frequency of seed cause it embeds commercial exchanges in a
transactions. Farmers do not keep records of web of obligations and holds the seller’s net-
such transactions, and estimates must rely on work hostage to appropriate role performance
the memory of those interviewed. In the tracer in the economic transaction. Also of interest
study, seed transactions involving current culti- here, is their point that exchange frequency re-
vars were registered, noting the year they took duces the extent of within-network exchanges—
place and allowing farmers to go as far back that is, more common in not-so-frequent acqui-
in time as desired. Findings indicated that sitions/transactions. Farmers obtained seed
recent transactions are more likely to be from many types of seed providers (e.g., family
remembered than those from the distant past. members, compadres, 8 neighbors, friends,
Notwithstanding these limitations, an estimate acquaintances, strangers, and others). In gen-
of the frequency of seed transactions was calcu- eral, the large majority of seed transactions
lated for the three most recent years and take place between people who know each
showed that on average acquisitions occur other prior to the seed transaction, and who of-
0.31 times per year per farmer and distributions ten share a feeling of social obligation toward
0.39 times per year per farmer—in both cases, each other (e.g., family members alone made
approximately once every three years (Badstue up 47.5% of seed providers in the seed flow tra-
et al., 2006). In 2001 only 24.2% and 20.9% of cer study).
farmers in the tracer study engaged in seed Particular types of transaction are not re-
acquisitions and distributions, respectively. In stricted to any one category of seed providers.
summary, seed transactions are relatively infre- Nevertheless, it appears that close social rela-
quent and do not involve a large number of tions between the seed provider and receiver
farmers every year. improve the latter’s chances of preferential
treatment, for example, in the type of transac-
(a) Embedded seed transactions tion or with regards to its terms or rates (Bads-
tue et al., 2003). The transactions that mediate
A seed transaction is an economic practice seed flows between farmers in the study com-
where a good is exchanged between two parties. munities, and the social relations associated
1586 WORLD DEVELOPMENT

with them, have been thoroughly analyzed (c) lack of sufficient seed for planting, and (d)
(Badstue et al., 2006). initiative by other farmers. Grouped by these
Trust is a key issue in seed transactions themes, Table 3 presents the distribution of
(Almekinders et al., 1994; Seboka & Deressa, informants’ reasons for acquiring seed from
2000). This is directly related to the lack of other sources, and their percentage relative to
transparency of seed. In addition, farmers the total number of seed acquisitions recorded
pointed out, that they generally prefer seed pro- in the seed flow tracer study.
viders who are easy to approach and believed Like farmers elsewhere, many farmers in the
to be willing to grant one’s request, especially Central Valleys are curious and eager to learn
if one cannot pay for the seed with money and explore new options. While they may be
and therefore depends on negotiating another well aware that a maize variety that works for
type of transaction. Finally, the trustworthiness others may not work for them, they also recog-
of the seed receiver is relevant to seed provid- nize that the maize of others may have advan-
ers, for example, with regards to the types of tages or provide traits that may be worthwhile
seed transactions that involve ‘‘payment’’ having. Furthermore, many farmers in the
forms other than money, and where the seed study area believe that ‘‘foreign’’ seed can even-
provider depends on the seed receiver uphold- tually ‘‘acclimatize’’ to local conditions, if
ing his/her part of the deal. planted and selected under those conditions.
Additional research on farmers’ seed acquisi- These elements lead to many instances in which
tion strategies in the study area, indicates the farmers ‘‘try out’’ other materials they come
significance of differences in farmers’ trust in across, combine them or even cross them with
seed providers: Seed acquired from people the their own materials to ‘‘see if it works.’’ These
farmer knows and trusts is generally perceived different farmer experiments usually involve
as entailing less risk of crop failure due to inad- only small quantities of seed or land, thereby
equate seed, than seed acquired from unknown minimizing the risks related to experimenta-
or impersonal sources, such as market vendors tion.
or commercial seed traders (Badstue, 2004). In When new households start farming on their
other words, the more the seed receiver knows own account, they usually get seed from par-
and trusts the seed provider, the less the per- ents or other close relatives. Not surprisingly,
ceived risk or uncertainty related to incomplete therefore, this counts as an important reason
or incorrect information. for seed acquisition.
Finally, relations of trust are conducive to Lack of seed may be due to seed loss or to
easy access to trustworthy information at low not being able or willing to save sufficient seed.
costs. Farmers may already know the charac- Seed loss may occur because of low yield or
teristics of varieties used by kin or close friends, total harvest loss, due to drought, water log-
and they can easily obtain more information ging, insect attacks, weeds, hail, lodging, or
(Badstue, 2004; Badstue et al., 2006). Thus, in poor management. Seed may be lost during
accordance with DiMaggio and Louch (1998) storage due to insects or rodents. A farmer
findings mentioned above, acquiring seed from may not save seed, or at least not enough, be-
social relations of trust can be seen as a way of cause he or she had to sell or eat everything that
reducing the problem of lack of transparency in was harvested including the seed set aside, as a
seed. This, in turn, helps reduce farmers’ trans- result of insufficient production, an emergency,
action costs in relation to seed acquisition to a or a crisis. Farmers who produce maize for ani-
minimum (Badstue, 2004). mal feed may harvest before seed is produced.

(b) Reasons for acquiring seed

Table 3. Reasons for acquiring maize seed
Farmers reasons for acquiring seed were Themea No. %
identified and described during the focus group Experimentation 100 32.5
discussions and in-depth interviews with key Commencing to farm 86 27.9
informants, and were later quantified during Lack of sufficient seed 56 18.2
the tracer study. Based on farmers’ own classi- Other farmer’s initiative 17 5.5
fications, the reasons provided by the infor- Others 49 15.9
mants for acquiring maize seed from other
sources can be divided into four main themes: Total 308 100
a
(a) experimentation, (b) commence farming, Unknown for 18 of the 326 seed acquisitions.
 THE DYNAMICS OF FARMERS’ MAIZE SEED SUPPLY PRACTICES 1587

Obviously, seed loss may also occur as the re- Table 4. Reasons for distributing seed
sult of several converging factors. People, who Themea No. %
for some reason decide not to plant maize for
some time, for example, due to temporary Help the recipient 131 69.3
Obtain something in return 50 26.5
migration, face a similar situation when they
Others 8 4.2
take up planting again, due to the relatively fast
decline in maize seed germination rate and Total 189 100
vigor (Morris et al., 1998). a
Unknown for 1 of the 190 seed distributions.
Farmers sometimes receive seed from other
farmers without having asked for it, for exam-
ple, when they agree to another farmers’ re- tion of being a ‘‘good farmer’’ is the idea that
quest for a seed-for-seed exchange. Even if a one should not refuse to help a fellow farmer
farmer has not actively looked for the seed, asking for seed if one has sufficient seed to
he or she may eventually decide to plant it, share. As stated by various informants: ‘‘I gave
although this does not always happen. Also, him the seed because I had it!’’
farmers sometimes receive small amounts of The other theme involves obtaining some-
seed as gifts. For example, one farmer’s sister, thing in return for the seed, mostly cash and
who lives in another town, each year, brings also seed. It is important to note that most dis-
small amounts of seed from her own maize tributions with the purpose of obtaining money
field, when she comes to visit. Her brother were associated with only two persons who are
plants this seed and explains that he regards it known to sell seed every year as a way of sup-
as a token of the affection between his sister plementing their income. As described in the
and himself and as a way to stay ‘‘close,’’ in next section, most acquisitions were purchases,
spite of the distance that separates them. In but relatively few seed providers were moti-
any case, these reasons for acquiring seed are vated exclusively by the view to obtaining
relatively infrequent. money in return. This, in turn, suggests that
In many cases, seed loss appears to be associ- the primary motive for farmer-to-farmer seed
ated with a certain social stigma, even though distribution rarely is to generate a profit. These
the cause for seed loss may be beyond the farm- findings suggest that there is a strong cultural
er’s control. Informants explained that seed value in the study area associated with being
loss sometimes is associated with laziness, lack helpful to others, as long as one is able to do
of knowledge, and inappropriate working prac- so while covering one’s own needs. For exam-
tices, etc. Meanwhile, never to have lost one’s ple, people who are known to have plenty of
seed is a cause for pride for many farmers. seed, but who are nevertheless not willing to
Obviously, these circumstances do not motivate provide seed to others, are thought of as selfish.
people to talk about the occasions on which This links with data from the in-depth inter-
they may have lost their seed, and it is possible views with key informants, which showed that
that this influenced the answers to the tracer an important motivating factor for many seed
study. providers is that the person requesting the seed
has a genuine need for it. Finally, it should be
(c) Reasons for distributing seed mentioned that this seems also to be part of a
common sense of reciprocity; as one of the
The flip side of acquiring seed is distributing informants pointed out: ‘‘What goes around,
it. The reasons provided by farmers for distrib- comes around.’’ On the other hand, the fre-
uting seed to other farmers can be divided into quency of purchase as transaction type and
two main themes: (a) to help the recipient and the above-mentioned broad willingness to sup-
(b) to obtain something in return, for example, ply seed to a buyer, also suggest that monetary
money or seed. Table 4 shows how the seed dis- gain could often be part of the motive for sup-
tributions recorded in the tracer study distrib- plying seed.
uted across these themes.
In another paper (Badstue et al., 2006), we
have argued that access to seed in the study 6. DISCUSSION
area may be conceptualized as part of a general
social responsibility for mutual assistance. A series of factors that influence farmers’
Most seed providers stated that they distributed seed supply practices have been identified. In
seed to help the seed receiver. Linked to the no- relation to problems regarding seed supply
1588 WORLD DEVELOPMENT

farmers in the study area negotiate solutions on the advantage that one knows the seed was pro-
an ad hoc basis. However, considerable varia- duced in that community, and therefore is
tion exists between farmers, even with regards likely to be adapted to local agro-ecological
to individual farmers who on different occa- conditions. Even if environmental conditions
sions may respond differently to seemingly sim- vary within the same community, in most cases,
ilar problems. the farmer would easily be able to determine
The central axis of the seed system in the the likelihood that the seed will be adapted to
study communities appears to be farmers’ prac- the conditions of his/her own land.
tice of selecting and saving seed from one year Finally, using social networks to acquire seed
to another. This is the source of seed for the is effective because it embeds the seed transac-
large majority of maize area planted in this re- tions in a web of obligations and, as pointed
gion, and for the individual farmer this practice out by DiMaggio and Louch (1998), ‘‘holds
can help reduce perceived risk and costs. It is of the seller’s network hostage to appropriate role
further symbolic importance for some, who performance.’’ Thus, acquiring seed via one’s
take pride in being self-sufficient in seed or re- social network can be seen as a way of reducing
gard the family seed as something valuable they the risk of planting inappropriate seed, that is,
have in trust and must pass on to subsequent maize that does not correspond to one’s pro-
generations. In addition, the common practice duction or consumption objectives, or which
of saving seed is a vital element in maintaining is not adapted to the local environmental con-
seed security at community level. The wide- ditions.
spread practice of saving enough seed for the The notion of the ‘‘good farmer’’ may also
next planting, and some extra for any contin- come into play in relation to maize seed trans-
gencies, provides a buffer against seed loss at actions. As mentioned above, it is thought
the household level, and also helps ensure that, appropriate ‘‘good-farmer-behavior’’ to help
in general, seed can be obtained locally when other farmers in need, when possible and within
needed. reason. In as far as a farmer can spare the seed,
During seed selection farmers exercise selec- this includes acting as seed provider on the
tion pressure in an attempt to enhance favored request of other farmers who need seed. This
varietal traits and lessen the influence of unde- sense of social responsibility linked to the no-
sired traits. Analysis of the genetic structure of tion of ‘‘a good farmer’’ may well be triggered
maize landraces collected in the study commu- when a request for seed is brought forward.
nities has shown a strong structure associated Meanwhile, failing to save seed is sometimes
with farmers and communities, when pheno- associated with a certain disgrace or loss of
typic traits are analyzed (Pressoir & Berthaud, prestige. While this may play a role as an incen-
2004). Structure on phenotypic traits indicate tive for farmers to live up to this standard, it
that varieties collected from the same farmer may also play a role in reducing the problem
or same community are more similar in their of free riders. 9
characteristics—mainly ear and grain traits— The practices and dynamics that make-up the
than those that were collected from other farm- local seed system in the study area appear to be
ers or other villages. This indicates that human grounded in a set of shared views and condi-
selection is playing a key role in creating and tions, which in themselves are based on the
maintaining different types of maize, and hence, agro-ecological, cultural, and social environ-
phenotypic diversity (Perales et al., 2005; Press- ments in which these farmers operate. Local
oir & Berthaud, 2004). seed supply in these communities is not based
The problem of nontransparency of seed and primarily on commercial motives. It is mainly
the issue of genotype-by-environment interac- part of a moral system based on trust and social
tion entails certain fundamental problems, responsibility.
which mean that acquiring maize seed is not a It should be mentioned that once in a while a
trivial transaction. farmer may acquire seed at the regional market
In most cases farmers’ easiest source of or elsewhere outside the community in order to
knowledge and trustworthy information about deliberately avoid the various implications that
maize and maize seed, as well as their preferred may arise from acquiring seed from other farm-
source of seed, is people they know and trust, ers in the community; such as expectations of
who in many cases also farm in the same com- reciprocity and the feeling of ‘‘indebtedness,’’
munity. Furthermore, acquiring seed from an- the ‘‘stigma’’ of having lost seed, etc. Likewise,
other farmer from the same community has it should be noted, that while the types of
 THE DYNAMICS OF FARMERS’ MAIZE SEED SUPPLY PRACTICES 1589

transactions not involving money payments study area. Farmers’ practice of saving and
may be attractive under some circumstances, selecting seed both constitutes the basis of the
under other circumstances a farmer may find phenotypic diversity observed in the study area
that paying for the seed with money can pro- and the foundation of the seed system. One
vide a swift and less personal option, and thus could say that each farmer is creating and
be preferable. maintaining his/her own unique maize varie-
When acquiring maize seed from a stranger, ties. Second, gene flow is important to bring
for example, at a regional market place, there new traits and modify varieties to fit farmers’
are no means of knowing its genotype-by-envi- needs—as farmers do when they experiment
ronment adaptation or other characteristics with ‘‘foreign’’ seeds and mix them with their
apart from what the vendor claims. As several own. Third, gene flow may also be important
farmers exclaimed when referring to commer- to maintain the viability of these landraces in
cial traders: ‘‘They just want to sell their the face of deleterious mutations, or simply to
goods!’’ In general, acquiring seed from un- avoid inbreeding depression. The current seed
known sources is perceived by farmers as system allows farmers to continue this process
entailing a certain risk of acquiring inappropri- of experimentation and incorporation of new
ate seed (Badstue, 2004). varieties or traits into their repertoire.
Recognizing that other maize varieties may Given the limited, relative frequency of seed
be useful or contain desirable characteristics, loss in the study area, farmers current seed sup-
farmers experiment with and ‘‘try out’’ seed ply practices appear relatively efficient in terms
of other kinds of maize than their own. This al- of maintaining local crop genetic resource
lows farmers to see for themselves the traits and diversity. While the system depends on suffi-
performance of the maize variety in question cient opportunities for obtaining seed from oth-
and judge whether it is appropriate for their ers when the need arises, at the moment this
individual needs and preferences. Meanwhile, does not appear to be a major limitation. From
farmers in the study communities do not asso- a population genetic point of view the seed sys-
ciate small-scale experimentation with signifi- tem appears to work well and be efficient in
cant costs, for example, in terms of extra time continuing to maintain a diversity of maize
or labor (Badstue, 2004). The principal cost is landraces and contribute to the conservation
the risk that the experiment will not be success- of maize genetic diversity (Bellon et al., 2003;
ful. However, this is manageable due to the Smale et al., 2003).
small scale of most farmer experiments, which
reduce the risk of major crop failure. Farmer
experiments therefore serve both the purpose 7. CONCLUSIONS
of information and of risk control. In addition
these experiments are also used to multiply In this paper we have presented an analysis of
seed. If the farmer decides to incorporate the the dynamics of smallholders’ maize seed sup-
‘‘new’’ varieties into the household’s maize rep- ply practices in the Central Valleys of Oaxaca.
ertoire, or alternatively mix it with seed of their The core principle of the local seed system in
own varieties in order to create new, desirable the study area is farmers’ reliance on selecting
combinations, they may therefore not need to and saving seed from the previous harvest.
acquire seed again. Farmer-to-farmer seed flow is an occasional
Pressoir and Berthaud’s (2004) research on event, which mostly involves relatively small
the genetic structure of landraces collected in quantities of seed, yet, it remains an essential
the same study communities has shown absence element of the seed system in order to ensure
of structure in these populations when neutral local seed security and maintain local maize
markers are analyzed. By definition neutral genetic diversity.
markers are not under selection. They provide Even though farmers in the study communi-
information on the evolutionary history of a ties live and farm under seemingly similar con-
population, that is, migration, bottlenecks, ditions, individual preferences, production
drift. The fact that no structure was found indi- conditions, and production objectives may vary
cates that migration (gene flow) among these considerably from one farming household to
populations has been strong enough to com- another. In this context, farmers’ widely ob-
pensate for the effects of bottlenecks and drift. served practice of selecting and saving seed
The results on genetic diversity complement from the previous harvest provides the basis
the present analysis of the seed system in the for a pool of considerably diverse and locally
1590 WORLD DEVELOPMENT

adapted maize varieties. Furthermore, at any initiatives related to introduction of improved

one time the number of farmers in need of, or varieties.
requesting seed from others, is relatively lim- The importance of farmer demand for differ-
ited. This, in turn, makes it relatively easy for ent maize types should be considered and the
these farmers to find someone who can provide germplasm involved should be assessed in rela-
the seed they need at a relatively low cost and tion to local production and consumption pref-
risk. erences, as well as local production conditions.
In combination with the occasional introduc- Farmers demand for seed of other varieties is
tion of new maize genetic material from other relatively infrequent, and mostly involve rela-
farmers or elsewhere, the on-farm reproduction tively small amounts of seed. Under these
and selection of seed of preferred varieties circumstances supplying seed may not be a prof-
provides a relatively efficient basis for the main- itable enterprise. Interventions such as those
tenance of local maize genetic diversity. practiced in the CIMMYT/INIFAP research
Whereas no imminent threat to maize genetic project demonstrated farmers’ interest in acquir-
diversity in the study area was identified in the ing seed of other maize varieties. However, the
research referred to here, the important role of average quantity of seed per acquisition was just
gene flow in preserving the viability of farmer 4.3 kg, which may not be a problem if the de-
varieties in the area was pointed out. mand is only for one or two varieties. However,
Furthermore, the current seed system pro- if the demand is for many different varieties,
vides a relatively secure seed supply of a range the costs of producing, managing, and selling
of valued farmer varieties. However, as men- seed may be too high for a commercial provider,
tioned, the farming systems in the study area since commercial seed enterprises most likely
are characterized by low productivity and have to supply larger quantities of fewer varieties
maize production per area is low. The use of to be financially viable.
improved seed could be one option for increas- Maize continues to play an important role in
ing maize productivity. It is possible that this the study area with regards to food security.
could help farmers produce sufficient maize While farmers in these communities are often
on less land, thereby freeing up land for other, curious and interested in trying out new things,
more income generating, purposes, or for in- in many cases, they are also concerned about
creased maize production. risks in relation to their maize production.
Finally, the current seed system is efficient in Unstable seed supply, timeliness of seed supply,
supplying farmers with maize seed of varieties higher costs of improved seed—in the study
that correspond well to local consumption area approximately 5–7 times the cost of farmer
and production objectives, and which are able saved seed—all add to the more generic prob-
to perform under local agro-ecological condi- lem of nontransparency of seed, and can influ-
tions. In the few cases where improved varieties ence and maybe hamper the adoption process.
are used in the study communities it is mainly Under such circumstances one may likely ex-
for corn-on-the-cob and feed production, not pect farmers first to experiment for some time
for the production of grain for consumption. with small quantities of seed, before deciding
This suggests that the improved varieties avail- whether or not to adopt. Emphasis should
able in the study area may be considered to therefore be on medium to long-term interven-
have inferior consumption characteristics from tions.
the point of view of the local population. Farmer skepticism with regards to the trust-
worthiness of market traders and commercial
seed vendors relates to information quality
8. IMPLICATIONS FOR POLICY about seed traits and quality. However, it seems
CONSIDERATIONS reasonable to think that it would also relate to
additional aspects such as the stability, reliabil-
From a policy point of view, the analysis of ity, and timeliness of seed supply. Obviously,
the dynamics of farmers’ maize seed supply improving confidence in commercial seed trad-
practices in the Central Valleys of Oaxaca ers presents a challenge.
raises several relevant issues. Findings point To a large extent many of these challenges
to both challenges and opportunities and may lead back to the fundamental problem of non-
be relevant in a broader perspective, for exam- transparency of seed, and to the importance
ple, in relation to commercial seed sector devel- of trustworthy information about maize seed
opment in this area of Mexico, or for other of different varieties. Identifying ways of
 THE DYNAMICS OF FARMERS’ MAIZE SEED SUPPLY PRACTICES 1591

conveying the relevant information to the users products through local channels that farmers
of the seed in a straight-forward and trustwor- felt comfortable with and could easily relate to.
thy way would seem a useful contribution in In the study reported on here, no specialized
this regard. seed-focused institutions of collective action
Despite a strong concern for risk avoidance, were identified. Rather than maintaining spe-
farmer interest in and willingness to experiment cialized networks for seed needs, which occur
and learn about new and different maize varie- relatively infrequent, farmers tend to ‘‘piggy-
ties, presents an opportunity for the intro- back’’ seed needs on other networks of social
duction of improved varieties. Though this relations on an ad hoc basis. The problem of
experimentation mostly involves small quanti- nontransparency of seed is one of the factors,
ties of seed, it nevertheless presents a window which may influence the decision to transact
of opportunity for anyone who would like to with a friend or a relative, as a response to
introduce alternative maize germplasm. the perceived risk or uncertainty this fosters.
Many farmers in the study communities ex- Under these circumstances, development inter-
press an attitude of generalized trust in other ventions at the community level, whether direc-
farmers, and farmer-to-farmer information ted toward conservation or introduction of
flows can play an important role in relation to improved seed, should focus on existing social
the spread of information regarding innova- organizations rather than trying to create new
tions and new technologies (e.g., Ryan & organizations dedicated to seed supply.
Gross, 1943). This may also be regarded as an Finally, in relation to the conservation of
opportunity in relation to the introduction of crop genetic diversity on-farm, the study points
improved varieties, and can be used actively. to the importance of sustaining seed flows be-
The concept of ‘‘farmer–dealers’’ was, for tween farmers. In this regard, seed fairs and
example, an important element in the spread other interventions that promote knowledge-
of hybrid maize in the US (Duvick, 1998). By based information flows and access to ‘‘new’’
using local farmers as their agents, seed compa- interesting varieties for experimentation at low
nies and government extension promoted their cost and low risk, represent interesting options.

NOTES

1. By seed system we refer to the set of sources of seed transaction costs in relation to seed acquisition (Badstue,
and related information, practices and transactional 2004). Both of these research projects built on previous
arrangements on which farmers rely to obtain seed for research undertaken by the International Maize and
agricultural production. Wheat Improvement Center (CIMMYT) and Instituto
Nacional de Investigaciones Forestales Agrı́colas y
2. In another very illustrating example of this, Perales, Pecuarias (INIFAP) during 1997–2002 in the same
Benz, and Brush (2005) have shown that ethnolinguistic region (Bellon et al., 2003; Smale, Aguirre, Bellon,
differences between ethnic groups in Chiapas could Mendoza, & Manuel Rosas, 1999).
explain the morphological and agricultural differences
found between varieties, while neutral markers showed 4. A Microsoft Access database was used for managing
no genetic differentiation. This means that genetic the data.
exchanges, that is, gene flow, had been sufficient over
time to eliminate genetic differentiation. 5. For more on genotype-by-environment interaction,
see for example, Bänziger and Cooper (2001).
3. The analysis presented here draws on findings from
three different but interrelated research projects carried
6. As an open-pollinated crop, maize is subject to
out in the same area of the Central Valleys of Oaxaca,
cross-pollination (Morris, 1999).
Mexico. The major part of this paper is based on a
project which examined the role of collective action in
the conservation of maize genetic diversity on-farm 7. An almud is a commonly used volume measurement
(Badstue et al., 2005). In addition, the paper draws on for grain or seed in the Central Valleys (Smale et al.,
another study focusing on the identification of farmers’ 1999). One almud of maize is approximately 4 kg.
1592 WORLD DEVELOPMENT

8. From the word compadrazgo, referring to a ritual 9. The fact that failing to save seed is associated with a
kinship, somewhat similar to the relation known else- certain loss of prestige may deter certain persons from
where as godparents, through which close relations of just asking others for seed instead of going through the
loyalty, mutual help, reciprocity, and confidence are trouble themselves of selecting and saving seed.
established and formalized (Cordero Avendaño de
Durand, 1997).

REFERENCES

Aguirre Gómez, J. A. (1999). Análisis regional de la ico. Genetic Resources and Crop Evolution, 50(4),
diversidad del maı́z en el Sureste de Guanajuato. 401–416.
PhD disertación. Universidad Nacional Autónoma Bellon, M., & Brush, S. B. (1994). Keepers of maize in
de México, México, DF. Chiapas, Mexico. Economic Botany, 48(2), 196–209.
Almekinders, C. (2001). Management of crop genetic Cordero Avendaño de Durand, C. (1997). La vara de
diversity at community level. Germany, GTZ: Esch- mando. Costumbre jurı́dica en la transmisión de
born. poderes. Oaxaca de Juárez, México: Biblioteca del
Almekinders, C., & de Boef, W. (2000). Encouraging 465 Aniversario.
diversity: The conservation and development of plant Cromwell, E. (1990). Seed diffusion mechanisms in small
genetic resources. London: Intermediate Publications farmer communities: Lessons from Asia, Africa, and
Ltd. Latin America. ODI Network Paper 21. London:
Almekinders, C. J. M., Louwaars, N. P., & de Bruijn, G. Overseas Development Institute.
H. (1994). Local seed systems and their importance DiMaggio, P., & Louch, H. (1998). Socially embedded
for an improved seed supply in developing countries. consumer transactions: For what kinds of purchases
Euphytica, 78(3), 207–216. do people most often use networks? American
Badstue, L. B. (2004). Identifying the factors that Sociological Review, 63(5), 619–637.
influence small-scale farmers’ transaction costs in Duvick, D. N. (1998). The United States. In M. Morris
relation to seed acquisition. An ethnographic case (Ed.), Maize seed industries in developing countries
study of maize growing smallholders in the Central (pp. 193–212). Boulder, CO: Lynne Rienner.
Valleys of Oaxaca, Mexico. ESA Working Paper No. Granovetter, M. (2005). The impact of social structure
04-16. Rome: Food and Agriculture Organization of on economic outcomes. Journal of Economic Per-
the United Nations (FAO). spectives, 19(1), 33–50.
Badstue, L. B., Bellon, M. R., Berthaud, J., Juárez, X., Gudeman, S. (2001). The Anthropology of Economy.
Manuel Rosas, I., & Solano, A. M. (2006). Exam- Oxford: Blackwell Publishing Ltd.
ining the role of collective action in an informal seed Hernandez, E. (1985). Maize and man in the greater
system: A case study from the Central Valleys of Southwest. Economic Botany, 39(4), 416–430.
Oaxaca, Mexico. Human Ecology, 34(2), 249–273. INEGI (Instituto Nacional de Estadı́stica, Geografı́a e
Badstue, L. B., Bellon, M. R., Berthaud, J., Ramirez, A., Informática) (2001a). XII Censo General de Pobla-
Flores, D., & Juarez, X. (2005). Collective action for ción y Vivienda 2000. Mexico: INEGI.
the conservation of on-farm genetic diversity in a INEGI (Instituto Nacional de Estadı́stica, Geografı́a e
center of crop diversity: An assessment of the role of Informática) (2001b). Aspectos Geográficos de Oax-
traditional farmer’s networks. CAPRI Working aca. Mapa de Temperatura Media Anual. <http://
Paper, no. 38. Washington: International Food oax.inegi.gob.mx/territorio/espanol/temperat.html>
Policy Research Institute (IFPRI). (last update: 29/03/2001, 13:04).
Badstue, L. B., Bellon, M. R., Juárez, X., Manuel Rosas, Jarvis, D. I., Myer, L., Klemick, H., Guarino, L., Smale,
I., & Solano, A. M. (2003). Social relations and seed M., Brown, A. H. D., et al. (2000). A training guide
transactions among small-scale maize farmers: A for in-situ conservation on-farm. Version 1. Rome:
case study from the Central Valleys of Oaxaca, International Plant Genetic Resources Institute
Mexico. Economics Working Paper 02–02. Mexico (IPGRI).
City: International Maize and Wheat Improvement Long, N. (2001). Development sociology. Actor perspec-
Center (CIMMYT). tives. London: Routledge.
Bänziger, M., & Cooper, M. (2001). Breeding for low Louette, D., Charrier, A., & Berthaud, J. (1997). In situ
input conditions and consequences for participatory conservation of maize in Mexico: Genetic diversity
plant breeding: Examples from tropical maize and and maize seed management in a traditional com-
wheat. Euphytica, 122(3), 503–519. munity. Economic Botany, 51(1), 20–38.
Bellon, M. (2004). Conceptualizing interventions to Matsuoka, Y., Vigouroux, Y., Goodman, M. M.,
support on-farm genetic resource conservation. Sanchez, J., Buckler, G. E., & Doebley, J. (2002).
World Development, 32(1), 159–172. A single domestication for maize shown by multilo-
Bellon, M., Berthaud, J., Smale, M., Aguirre, J. A., cus microsatellite genotyping. Proceedings of the
Taba, S., & Aragón, F. (2003). Participatory National Academy of Sciences, 99(9), 6080–6084.
landrace selection for on farm conservation: An Morris, M. (1999). Maize in the developing world:
example from the Central Valleys of Oaxaca, Mex- Waiting for a green revolution. In M. Morris (Ed.),
 THE DYNAMICS OF FARMERS’ MAIZE SEED SUPPLY PRACTICES 1593

Maize seed industries in developing countries Seboka, B., & Deressa, A. (2000). Validating farmers’
(pp. 35–54). Boulder, CO: Lynne Rienner. indigenous social networks for local seed supply in
Morris, M., & López-Pereira, M. A. (1999). Impacts of Central Rift Valley of Ethiopia. Journal of Agricul-
maize breeding research in Latin America 1966–1997. tural Education and Extension, 6(4), 245–254.
Mexico, DF: International Maize and Wheat Smale, M., Aguirre, A., Bellon, M., Mendoza, J., &
Improvement Center (CIMMYT). Manuel Rosas, I. (1999). Farmer management of
Morris, M., Rusike, J., & Smale, M. (1998). Maize seed maize diversity in the Central Valleys of Oaxaca,
industries: A conceptual framework. In M. Morris Mexico: CIMMYT-INIFAP. 1998 Baseline socio-
(Ed.), Maize seed industries in developing countries economic survey. CIMMYT Economics Working
(pp. 35–54). Boulder, CO: Lynne Rienner. Paper 99-09. International Maize and Wheat
Orlove, B. S., & Brush, S. B. (1996). Anthropology and Improvement Center (CIMMYT), Mexico DF.
the conservation of biodiversity. Annual Review of Smale, M., Bellon, M., Aguirre, A., Rosas, I. M.,
Anthropology, 25, 329–352. Mendoza, J., Solano, A. M., et al. (2003). The
Perales, H. R., Benz, B. F., & Brush, S. B. (2005). Maize economic costs and benefits of a participatory project
diversity and ethnolinguistic diversity in Chiapas, to conserve maize landraces on farms in Oaxaca,
Mexico. Proceedings of the National Academy of Mexico. Agricultural Economics, 29(3), 265–275.
Sciences, 102(3), 949–954. Sperling, L., Heidegger, U., & Buruchara, R. (1995).
Perales-Rivera, H. R., Brush, S. B., & Qualset, C. (2003). Enhancing small farm seed systems: Principles
Dynamic management of maize landraces in central derived from bean research in The Great Lakes
Mexico. Economic Botany, 57(1), 21–34. Region. Network on Bean Research in Africa,
Piperno, D. R., & Flannery, K. V. (2001). The earliest Occasional Publications Series, No. 15, Cali: Inter-
archaeological maize (Zea mays L.) from highland national Centre for Tropical Agriculture (CIAT).
Mexico: New accelerator mass spectrometry dates Subedi, A., Chaudhary, P., & Sthapit, B. (2003).
and their implications. Proceedings of the National Maintaining crop genetic diversity on-farm through
Academy of Sciences, 98(4), 2101–2103. farmers’ networks. In CIP-UPWARD: Conservation
Pressoir, G., & Berthaud, J. (2004). Population structure and sustainable use of agricultural biodiversity: A
and strong divergent selection shape phenotypic sourcebook (pp. 259–265). Philippines: International
diversification in maize landraces. Heredity, 92(2), Potato Center and Users’ Perspectives With Agri-
95–101. cultural Research and Development (UPWARD).
Ryan, B., & Gross, N. C. (1943). The diffusion of hybrid Swedberg, R., & Granovetter, M. (2001). Introduction
seed corn in two Iowa communities. Rural Sociology, to the second edition. In M. Granovetter, & R.
8(1–4), 15–24. Swedberg (Eds.), The sociology of economic life
Sanchez, J. J. G., Goodman, M. M., & Stuber, C. W. (pp. 1–30). Boulder: Westview Press.
(2000). Isoenzymatic and morphological diversity in Thiele, G. (1999). Informal potato seed systems in the
the races of maize in Mexico. Economic Botany, Andes: Why are they important and what should we
54(1), 43–59. do with them? World Development, 27(1), 83–99.