Between 2002 and 2006, I worked as a PhD researcher at Wageningen University. I conducted field work in the highlands of Guatemala to assess the temporal change and spatial distribution of maize diversity, as managed by farmers. The Guatemalan highland area is one of the areas of Latin America where ethnic identities are strong (roughly comparable to the Andes). The people in this area suffered one of the worst civil wars of the continent, which was most intensive in the early 1980s.
The basic idea of the dissertation was to find out how farmer seed management and exchange structure a crop genepool. Crop diversity, in this perspective, is not the result of a chain of unconnected decisions, but is shaped and maintained in a landscape of relations between farmers and their communities, the exchange of seeds and knowledge, crop growing environments, and farming practices.
I tried to work out the consequences of historical events on maize diversity in two papers. In a first paper, I elaborate a long-term perspective, starting in pre-colonial times. This is an unusual approach, and sometimes leads to somewhat speculative results, but gives also some remarkable insights and perspectives. I argue that the local orientation of highland communities will have caused rather locally specific patterns of crop diversity. I then try to show that astronomical population losses after the Spanish Conquest must have affected maize diversity, an argument that convinced at least one reviewer. I also put some question marks around the idea that the civil war of the 1980s would have caused an enormous decrease in diversity, contrary to the literature.
This last claim is elaborated in a second paper. After scanning through the historiographic and ethnographic literature, I came across two lists of maize varieties, drawn up by ethnographers in the early 20th century. One covered 27 townships in the highland region (Huehuetenango) including a township covered by the second list, Jacaltenango. For this area, the lists showed a strong agreement in the naming and classification of maize varieties. I then designed a simple survey to assess the changes, based on this unique baseline. The data showed that farmers did not feel dramatic changes in maize diversity had occurred. Most varieties were thought to be still present in the township. The data suggested, however, that new varieties and socio-economic change had not caused the whole-sale disappearance of the older varieties, but had nevertheless brought important changes in the sharing of knowledge about them. The war may have been an important factor in bringing about these changes by destroying the social fabric that supported the exchange of seeds and knowledge.
Farmer variety management is explored in a third paper. Based on a survey of a large number of households, this paper looks at short term decision making. Using a method called "classification trees" (advised by co-author Sytze de Bruin) I tried to tease out the factors that influence the decision to discard or introduce a variety. A clear trend in the data was that farmers tend to introduce short, quick-developing varieties from seed sellers, NGOs and the extension service. These varieties are reported, however, to loose their precocity over the years, making new introductions necessary.
A genetic analysis, written up in a fourth paper, tests some of the ideas developed in the previous papers. Based on a sample of maize seeds from one departamento in the highlands of Guatemala, it analyzes the environmental factors that correlate with genetic diversity, using multivariate statistics. The populations have significant spatial structure, and altitudinal differences explain a lot of the diversity. Locally, the mutual distance between the households is a good predictor of the genetic distance between the seed lots they have. Regionally, maize diversity is patterned in clusters. This indicates that local exchange between inmediate neighbours is important, but that regionally mutual distance plays a less important role than other factors. Over these longer distances other factors become important, such as centrality (of a market, for instance). Ear-related morphological variables (which have to do with race distinctions in maize and are easily observable to farmers), as well as adaptation (measured as yield) are also associated with phylogenetic distance. The analysis also showed the influence of improved varieties, as seed samples from the field closely associated with improved germplasm tended to be shorter in stature. The impact of improved germplasm was localised and constrained to certain places in the lower parts of the study area. This is clear evidence for the limited impact of previous breeding and seed marketing strategies.
The research has provided insights by focusing on the processes shaping the genepool from different, complementary angles. This has also allowed me to become acquainted with very different methods and styles of doing research. A next step in this line of research would be to do a region-wide genetic analysis of maize in order to test several important hypotheses formulated. In the postdoctoral research I am currently doing, I develop the analytical tools needed for this. I hope that one day, I will be in the position to apply these tools to Guatemalan maize, too.
I argue in the conclusions of my dissertation that maize improvement and seed delivery should be improved in the light of the findings of my research. There is a clear need for maize seed-based innovation, and the professional breeding activities of the past need to be expanded to include more diverse goals and to be linked more intensively to farming practice. Teaching individual farmers to select maize may not be very effective, however, as selection is labour-intensive, takes a long period to give visible results and has little value if the seed reaches only a few neighbours. It is argued that some specialization is needed, although the breeding strategies should allow for diversity. Regional sales points should be established to make seed marketing viable and innovation worthwile. Geo-informatics should be used to target seed innovation and marketing to specific geographical areas.
Links to all the published articles are provided above in the text, and can also be accessed from the Publications page. Some of them are behind a paywall, but the dissertation includes versions of all of them. The dissertation can be downloaded here. The introduction and conclusions were not published elsewhere.
