Conservation of Genetic Diversity and Improvement of Crop Production in Mexico: A Farmer-based Approach Project MILPA1 A project sponsored in large part by The McKnight Foundation Collaborative Crop Research Program 1997 Annual Report [Results in this report are preliminary and will be substantiated by further work and interpretation. The report should not be quoted without consultation and approval from the investigators. For details, please contact either the Mexico Management Office: R. Bye, Jardín Botánico, Instituto de Biología, UNAM, Ap. 70614, C.P. 04510, México D.F. MÉXICO (mcknight@ibiología.unam.mx) or the US Management Office: C.O. Qualset, Genetic Resources Conservation Program, University of California, One Shields Ave., Davis CA 95616-8602 USA (firstname.lastname@example.org).] I. Abstract/Summary In its third year, the MILPA project has made substantial progress towards reaching its initial objectives. The initial data-gathering phase has ended for the research on the economic and biodiversity background in the two main study regions of Mexico, the Chalco-Amecameca-Cuautla and the Sierra Norte de Puebla areas. Further fieldwork has benefited from lessons gained early in the project. Participatory selection and crop improvement have become primary activities and the first data from this work were obtained in 1998. The MILPA project has completed initial experiments on wild / cultivated gene flow and completed two cycles of on-farm selection for maize yield improvement. Introgression of genes from landraces into productive bean varieties has been initiated. The training program has maintained the desired level with 14 graduate students now participating in the project. Two senior investigators from Mexico spent a few weeks or longer in the U.S. MILPA begins its third year with great expectations because the first results from the various field and laboratory studies will become available for analysis and discussion. II. Introduction The overall goal of the MILPA project is to combine two apparently contradictory objectives, namely in situ conservation and improvement of local landraces, in the milpa cropping system of Mexico. To this end, the project has, in a first stage, constituted sub-projects based on the major commodities grown in the milpa system—maize, bean, and squash, and quelites or edible greens—and on the socio-economic component. Some highlights of 1997: Fernando Castillo was invited by IPGRI to give a presentation about MILPA in Rome. Rafael Pacska and Fernando Castillo were invited to join the advisory committee of the IPGRI in situ conservation project in Mexico. 1 Milpa: A traditional cropping system involving multiple uses of crops, especially maize, beans, and squash, in Mexico. MILPA: ―McKnight Integrated Landrace Preservation Activity‖. MILPA Page 2 Ramon Lastra, IPGRI Director for the Americas region has joined the MILPA Advisory Committee. The MILPA webpage has introduced reports, poster abstracts, and the project directory. Luis Eguiarte-Fruns, UNAM, has joined MILPA as supervisor to Ph.D. student S. Montes. June 97. Participation in the McKnight Foundation Conference of CCRP Projects, ―Cultivating Change: Global Partnerships for Food Security‖ Lake Tahoe, CA. Six posters and one oral presentation about MILPA were made. August 97. Paper presentation: Perales, Hugo, S.B. Brush, C. Qualset, J.E. Taylor 1997 Agronomic and economic competitiveness of maize landraces and in situ conservation in the Amecameca and Cuautla valleys of Mexico. Paper presented at CIMMYT conference on ―Building the Theoretical and Empirical Basis for the Economics of Genetic Diversity and Genetic Resource Conservation in Crop Plants,‖ Stanford, CA. September 97. Initial discussions to link CCRP project ―Andean Root and Tuber Crops‖ to MILPA. Marleni Ramirez from ART has joined the MILPA Advisory Committee and Steve Brush from MILPA has joined the ART Advisory Committee. October 97. Annual meeting for the second year of the grant; Colegio de Postgraduados, Montecillos, Mexico was attended by over 40 members of the project, representatives from the McKnight Oversight Committee, D. Duvick and P. Pengali, attended, as did MILPA Advisory CoChairs Francisco Cardenas and P. Bretting and other members of the MILPA Advisory Committee. The agenda included: Review of the year‘s results and plans for project renewal. Reports on integration among scientific disciplines in addressing the project‘s objectives in research and training. Student reports of research participation. Participation by Marleni Ramirez of the Andean Roots and Tubers project, for the purpose of planning joint activities. November 97. Six Investigator Scientists from Mexico and five from the US met at UC Davis for project planning and preparation of the project renewal proposal. September 96. – June 97 Sabbatical leave by Antonio Yunez (Colegio de Mexico) – UC Davis September 97– August 98. Sabbatical leave by Alfonso Delgado (UNAM) to Montana State University, Bozeman III. Scientific achievements and future plans The overall goal of the project is to increase the self-sufficiency of local farmers and their communities in terms of plant genetic resources. We hope to enhance their ability to protect, use, and enhance local plant germplasm, probably in combination with nonlocal germplasm. The project has been focused on a major cropping system utilized in Mexico (as well as other parts of the world), consisting of maize, beans, squash, and edible greens MILPA Page 3 (or quelites). The three objectives established at the outset of this project have been followed organizationally and strategically to advance knowledge related to the project goal. This report, based on results obtained cumulatively from years 1 and 2, begins to show the accomplishments of the project, some of which are highlighted here. More detailed reports are available through the project management offices at UNAM, Mexico City and UC Davis, California. Study areas are mainly directed to two areas in Mexico (Fig. 1) known as the Chalco/Cuautla and Sierra Norte de Puebla (SNP) where biological and socio-economic approaches can be integrated. Other regions have been selected to meet certain needs of the project. Mexico City Chalco Amecameca Cuautla Sierra Norte de Puebla Figure 1. The Chalco-Amecameca-Cuautla and Sierra Norte de Puebla study regions MILPA Page 4 Objective 1. Description and analysis of the relationships between farmer knowledge, socio-economic factors, and genetic diversity in the milpa agroecosystem. Over the past year two main foci have been addressed: (1) competition between crops within the milpa and between milpa production and other activities that shapes crop genetic outcomes on individual farms, and (2) economic and crop genetic linkages among households within communities, which may shape crop genetic outcomes both on individual farms and on a larger (in this case, community-level) scale. A general analysis of valuation of natural resources lacking market prices—including crop genetic resources—was completed. Even though crop genetic resources often do not have a market price, most of them have an economic value. This is true for genetic resources of crops grown by traditional farmers in Mexico. The economic value of crop genetic resources within the MILPA, including quelite species, generally is ignored by economic research. This omission has consequences for the design of agricultural policies, which could otherwise act to safeguard this public good. Our assessment of the state of nonmarket economic valuation of natural resources in Mexico and other related issues appears in an article entitled ―La valuacion del ambiente y de los recursos naturales en la teoria y en la practica” (Yunez and Taylor, in-press). Chalco/Cuautla In the Chalco and Cuautla survey areas, two agro-climatic zones were studied, defined by altitude. Data from the full survey (119 households), maize sampling (98 households), and a variety comparison experiment (5 farms) have been processed and analyzed, leading to significant progress in completion of a Ph.D. dissertation by Hugo Perales (Colegio de Postgraduados) at UC Davis (Ecology). The main findings of the Chalco/Cuautla study are: 1. All communities under study have few (1 to 4) major maize types and several (5 to 10) minor types under cultivation, and all farmers have one or more of the major maize types. Over 80% of the area under cultivation with maize is allocated to the major varieties. Major maize types are only traditional varieties in the highland (2400 masl) and intermediate elevation communities (1700 masl), and both traditional and modern varieties at the lower elevation communities (1200 to 1400 masl). Between one-third and one-half of the farmers engage in cultivation of minor varieties, in all cases in small scale. Almost all minor maize types are traditional varieties. 2. There is no evidence that any of the major maize types have a higher yield and/or generate higher income than the others of the community, the only case with lower yield of one type is compensated by higher market price. 3. Farmer field trials compared traditional varieties and an institutionally recommended hybrid, subject to increasing plant densities and fertilizer rates, The traditional variety responded in a competitive manner with the modern variety under all conditions, and in a low-yielding site its yield was double that of the hybrid. 4. With the exception of the highland community, current major varieties are different from varieties cultivated 30 to 40 years ago in the communities studied. MILPA Page 5 5. Farmers are attentive to several characters of the varieties they cultivate, but all are yield related. 6. Opportunity costs of maintaining traditional varieties are not weighed against modern varieties but against other crop activities or income sources. Maize is not economically profitable in any of the communities studied and its maintenance as a crop reflects more a cultural priority than a profit-maximizing activity. 7. Origin of seed for each community is predominantly within the community, all communities present a small (<10%) amount of seed obtained from other communities; in the case of hybrids these are purchased each 2 to 4 years. 8. Change of major maize types within a community seems to take 10 to 20 years, minor types seem to change somewhat erratically. 9. Individual farmers are not particularly worried about losing their seed because they think new seed is easily available within the community. These findings suggest that in situ conservation of genetic diversity is widespread, although the processes of generating diversity and selection are maintained, rather than genetic integrity of a specific population. Distinct strategies are required for major and minor varieties. Since major varieties must keep a competitive edge, supporting in situ conservation of these varieties might imply improvement of farmer-based breeding techniques. It might not be a matter of supplying improved types of traditional varieties, but maintaining the large-scale involvement of farmers in breeding their crops. Due to the relative slow rate of change in adoption of major landrace maize types, those of interest could be easily monitored and appropriate actions can be taken, when required. The interest of many farmers of cultivating minor varieties could be supported by seed savers schemes, including mixing compatible types to reduce inbreeding problems. Sierra Norte de Puebla Socio-economic surveys were concluded in the two villages of Naupan and Reyesoghpan. Data were coded and entered into computer databases; they supported the community-level modeling discussed below. Through these surveys and site visits, we have now gathered information for Naupan and Reyesoghpan to learn the details of the milpa and staple production, as well as their socio-economic structure. For example, we know the types of corn that farmers plant and their use, the two villages‘ gross domestic products decomposed by agricultural and nonagricultural activities; the value added and distribution between factors of production; the land ownership structure; the income distribution between the villages‘ households; and the economic relations between the villages and the rest of the regions in which they are located. Social Accounting Matrixes (SAM) of the two villages, as well as SAM multiplier models were developed. With this basis we have performed some simulations, designed to measure the impact of ―exogenous‖ changes, such as the effects of land erosion on the villages‘ productive structure, income, and external linkages. We have developed a general theoretical model to address the interdisciplinary aspects of the milpa. The economic model of farmer behavior incorporates the cross-crop nature of the farmer‘s decision-making that characterizes the four crop groups within MILPA. It is also designed to examine competition from other cash crops that may be diverting farmer resources away from the milpa, which is seen in both Puebla field sites. MILPA Page 6 To date, the work of the socio-economic group has been centered on the household, based on household surveys. But the household is only the point of departure of our group‘s work. Households are components of rural communities. Interactions among households in communities may be critical in shaping biodiversity outcomes. In the past year, we have begun to add a community focus to the project, focusing on interactions among households that influence crop genetic resource diversity both directly (e.g., through exchange of crop germplasm) and indirectly (e.g., through market relationships that alter the incentives for planting traditional varieties, for polycropping within the milpa, etc.). Within a community there are many formal and informal relationships among households, and these relationships can significantly affect a household‘s cultivation decisions. For instance, one way to hedge against risk is to share it among households, as is done via various types of informal insurance schemes documented in Asia and some parts of Africa, or to diversify income production away from staples and into other crops, nonfarm production, or wage work, if it is available in the local economy. All of these may have profound implications for on-farm genetic diversity. Interdisciplinary Collaboration With stable footing in the research underway in Objectives 1 and 2, interdisciplinary collaboration has increased. In the Cuautla study region, the on-farm comparison between local and modern maize varieties helped to inform the farmer-based breeding program underway in the other study regions (SNP and Chalco). An October visit to bean research field sites and to quelite research in the SNP was undertaken to examine issues of farmer behavior that relate to the agronomic and ethnobotanic research. This on-site review was useful to show how data collected for crop diversity and gene flow, for example, are related to socio-economic research. For example, isozyme variation in maize samples collected during the socio-economic surveys in Chalco/Cuautla and SNP is now underway in the laboratory of Prof. Porfirio Ramirez (Colegio de Postgraduados, Montecillo). A bean survey was conducted among farmers (n=11) of the region based on the questionnaire developed the previous year with S. Brush and H. Perales. Some of the main results were: 1) most of the farmers do not buy bean seeds; 2) most farmers select seeds that are large and healthy, but color appears to play only minor role; 3) a majority of the farmers do not know how to differentiate morphologically wild and domesticated beans, although there are notable exceptions to this pattern; and 4) wild beans are generally not used (eaten) although previous generations may have done so before the construction of roads. Four on-farm bean trials, designed to discover farmer attitudes when confronted with new advanced breeding germplasm, were established near Amecameca at the edge of the Chalco area, but the farmers harvested them without the presence of the research group. Such disappointments are inevitable in on-farm trials. The following terminology for beans was ascertained: pixnet: trash bean (frijol de desperdicio); cinet: maize bean (frijol de maíz); milet: milpa bean (frijol de milpa); mecaet: climbing bean (frijol de bejuco o guía (mecate)); pitzahuaquet: slender bean (frijol delgado); xaco: mottled or striped bean without reference to color (frijol rayado, moteado o manchado, sin importar el color dominante); tacuahuaquet: hard bean (frijol duro, P. coccineus); exoyema: soft bean (frijol blando, P. polyanthus); acalete: (maize) stalk bean (frijol de caña (de maíz)). This wide range of terminology demonstrates the importance of beans in the local agriculture and provides a measure of the traits that matter to local farmers. MILPA Page 7 A survey instrument was prepared to learn about farmer knowledge and attitude about squash types and diversity in relation to the milpa cropping system was developed for a series of interviews in the Autlán Valley in the state of Jalisco. In this valley there is a rich diversity of squash types and long-term management of the domesticated types. The survey will be conducted in 1998. In connection with the studies of directed and natural gene flow between maize and its progenitor, teosinte, a survey of farmers‘ perception of teosinte was conducted. Teosinte commonly occurs in maize fields. Farmers surveyed in Jalisco responded as follows: 60% considered teosinte to be a useless weed, 34% use teosinte as a forage, and 6% believe that teosinte is important for maize breeding and some of them practiced some selection. In the same interviews 104 farmers were asked about the maize types which they grow. They reported 11 traditional varieties, each one being grown by 4 to 17% of the farmers, which accounted for 90% of the maize being grown. The remaining 10% were hybrid varieties. This corresponds to estimates of national production of 80% traditional and 20% improved or hybrid varieties. It also points to the limited diffusion of improved varieties into Mexican agriculture since there are more than 8 million hectares of maize planted each year. MILPA Page 8 Objective 2. Characterization of the structure of crop biodiversity and the magnitude of gene flow from wild or cultivated relatives to maize, bean, and squash Crop Biodiversity: Maize Collections of Landraces In the Chalco region 125 accessions were assembled from 28 farmer-communities. Hugo Perales collected 149 maize samples from three farmer-communities in the state of Morelos (Cuautla) and four in the State of México (Chalco) during his farmer-surveys. Representatives of maize populations from other areas with similar ecological conditions to the Chalco area were collected, including 12 samples from the states of Oaxaca and Michoacán and 70 accessions, reported as outstanding ones in previous studies or collected in areas with similar altitude and grown in the same way as the Chalqueño, from the states of Zacatecas, Durango, Guanajuato, Queretaro, Hidalgo, Michoacán, and Puebla were obtained from the genebank. In the Sierra Norte de Puebla, 173 accessions were assembled from 48 farmer-communities. During the socio-economic survey, Xochitl Juárez collected about 40 accessions from Reyeshogpan, SNP. Diversity Related to Culinary Uses Diversity analysis concentrated on several maize races from the Chalco area, with emphasis on Chalqueño. A clear preference for specific maize types for different uses (tortilla, tamales, pozole, and tlatloyo) was demonstrated. For example, Cacahuacintle is preferred for pozole preparation, ―palomo‖ for tamales, and blue maize for elotes. We evaluated 22 populations, taking samples of the typical cream dent Chalqueño, ―palomo‖, blue maize, dark colored kernel type, Cacahuacintle race, Ancho, and one that looks like Palomero Toluqueño, all of them collected in the Chalco region. Performance was assessed as raw materials for tortilla, tamal, and pozole. Principal component analysis with 36 traits (there is more information on similar or greater number of related traits, not presented in this report) showed that the three first principal components (PC) explained 56% of the global variation. PC-1 takes in account traits related to kernel density, physical or chemical components, and color in the positive direction, and tortilla, tamal and pozole yield related traits in the negative direction. PC-2 includes traits related with tortilla yield and volume yield of pozole in the positive direction, and tortilla quality and kernel size related traits in the negative direction. The plot of the first two principal components (Fig. 2) shows separated groups of the cream type accessions at the right hand side, the Cacahuacintle and the floury types in the opposite extreme, and the ―palomo‖ with the adapted Ancho in the middle. In terms of diversity utilization and conservation, it is important that different maize types showed different capabilities to produce specific meals. For genetic improvement it would be convenient to monitor grain yield and quality in terms of food processing performance. In a more aggressive strategy, separate breeding programs would be appropriate to develop improved varieties having better yield and better food processing quality, which may have an impact on higher market price. MILPA Page 9 PC-2 | | | | 7.5 + | | | R 5.0 + | | PT | A 2.5 + | Y | | A A C C 0.0 + A A C C C C | C | P xA | P C -2.5 + | | P | -5.0 + | CC | | -7.5 + | ---+----------+----------+----------+----------+----------+----------+--6 -4 -2 0 2 4 6 PC-1 Figure 2. Chalqueño maize populations scattered on the plane defined by the first and second principal components after characterization for tortilla, tamales, pozole, and physical and chemical kernel performance. C: typical cream colored kernels, A: blue floury, P: ―palomo‖ (white semi-floury), Y: yellow kernels, CC: Cacahuacintle, R: red floury, PT: Palomero Toluqueño, and xA: adapted Ancho. Grain Yield and Morphological Traits A particularly large effort has been made to characterize landrace maize diversity as related to performance. Maize accessions for these studies were collected from fields in 1995, some came from genebanks, and others were selected because of outstanding performance in previous studies. In total, of 144 accessions included in replicated trials in 1996 at three locations in the Chalco Amecameca area, 90 were new accessions. In 1997, 20 additional new accessions were included in the trials. The Chalqueño types performed very well, contributing to the fact that hybrids or improved varieties have not been adopted by farmers in this region. Yields, for example, at Poxtla were about 6.7 tons/ha in 1996 and 9.0 ton/ha in 1997, a more favorable year. Specifically, some landraces outyielded experimental hybrids and some newly collected landraces outperformed the same landrace collected some years earlier. Thus, it is obvious that genetic diversity in maize is dynamic, being landrace dependent reflecting the history of its evolution in the diverse maize production environments of Mexico. Diversity characterization and classification has also been done using traits related to reproductive structures and other attributes such as plant shape, phenology, etc. One factor MILPA Page 10 that makes these traits undesirable for taxonomic use is the large GxE interaction associated with them; thus the ranking of entries may change over locations. Preliminary trait assessment was run with information obtained in 1996. Combined analysis of variance and variance components were computed to obtain the ratio r = var(entries)/ [var (loc) + var(GxE)]. The best traits to assess diversity are those with small (GxE) variance and large r value. Out of 44 traits studied, kernel width, kernel shape, cob diameter, ear shape, and days to flowering appeared be the most useful attributes for diversity assessment. With the information on 53 traits recorded at the three locations in 1996, trait means for each accession over locations were taken for PC analysis. PC-1 and PC-2 explained 31 and 21 % of the overall variance, respectively. Important traits for PC-1 were days to flowering, plant and ear height, number of leaves, number of tassel branches and length of the branched portion of tassel in the positive direction, and disease ratings and the ratio plant height/ear height for the negative direction. For PC-2 kernel size, shape and weight; endosperm texture; internode length; and ear diameter contributed in the positive direction. The accessions on the plane defined by PC-1 and PC-2, shown in Figure 3, cluster the Chalqueño type at the upper central portion; from that point, to the right and downwards are accessions poorly adapted in the Chalco region, late maturing accessions with geographical origin located in Puebla (possibly chalqueño recombinants with tropical germplasm), late highland populations from Oaxaca and Michoacán (also with tropical germplasm). To the left and also downwards, there is a continuous variation towards the early highland materials. Seedling Vigor A most critical stage in the maize crop cycle is seed germination and seedling establishment. Farmers plant seeds quite deep (10 to 20 cm) to reach stored soil moisture. They will replant any failed hills. This leads to uneven growth throughout the field and potentially lower yields at the end of the season. Landraces have been faced with this cropping system for a long time and some of them may have more seedling vigor than others, particularly those coming from the most challenging environments. Using 15 half-sib families from each of 14 accessions, three series of experiments have been conducted to determine if these accessions differ in seedling vigor and if certain ones would be more desirable to farmers. Seedling vigor may be an important selection criteria for breeding programs. The studies included planting depth (10 and 20 cm in sand culture), osmotic stress (-12 bars, using polyethylene glycol in Petri dishes), and low temperature (12º C in rolled paper towels). The results of these three tests showed concordance for some accessions, especially for Palomo which showed poor performance, and four accessions which were consistently good in all types of test. Others, such as Ancho were good in low temperature, but not for osmotic stress or deep planting. Variation within the half-sib families indicated potential to select for greater environmental stress tolerance within a landrace type. A practical implication for on-farm crop improvement is that farmers could ‗over-plant‘ their hills with more seeds than usual and rogue out weaker seedlings one or two weeks after emergence. This form of mass selection may become part of the ‗farmer-breeder‘ protocol for on-farm maize improvement discussed in Objective 3. MILPA Page 11 PC-2 | | | | | | 5 + C C | C C C | CCCCCC C | MCCCCCC CCC | C CCCCMCCCC C | CM M CCC 0 + H C V MMMM C C | TH M P | PC PCM M H M | C m M O O | H H Z H C C C | P P G OC -5 + Z C | M M C C C | Z G C C | C C | M G | -10 + | | | P | | -15 + ---+----------+----------+----------+----------+----------+----------+--15 -10 -5 0 5 10 15 PC-1 NOTE: 42 obs hidden. Figure 3. Chalqueño maize accessions scattered on the plane defined by the first and second principal components after field characterization. Poxtla, Tlapala, and Tecamac, Edo. Méx. 1996. C ~ newly collected accession, M ~ accession from germplasm bank collected in the state of México, m ~ from the state of Michoacán, O ~ from the state of Oaxaca, P ~ from the state of Puebla, G ~ from Guanajuato, Z ~ from Zacatecas, H ~ from Hidalgo, and T ~ Tlaxcala. Principal components explain 31 and 21 % of overall variance, respectively. Crop Biodiversity: Bean In Zacatepec, State of Morelos, a set of 46 bean landraces collected by Hugo Perales were planted for characterization. Two groups were distinguished, one local and one introduced. In the introduced group, there were three tropical black-seeded types, probably introduced from the lowlands of Veracruz (Mesoamerican race), and six determinate bush lines that belong to the commercial class peruano/azufrado, probably introduced from the State of Sinaloa (Nueva Granada race). These introductions were mostly collected in the lower part of the transect Amecameca-Cuautla and were probably made available to farmers by merchants. In the local group, collected mostly at the higher elevations of the transect, most lines were climbers that were late to flower and mature (Jalisco race) and three belonged to MILPA Page 12 the ayocote group (P. coccineus). Crop Biodiversity: Squash Chilacayote, Cucurbita ficifolia There are two types of chilacayote (C. ficifolia), ―fino‖ and ―corriente‖. The fino type generally possesses a round fruit that averages about 1 kg. It is generally cultivated in monoculture for vegetable production. The corriente type is characterized by an elliptical fruit that averages about 7 kg and is generally grown as a maize intercrop in milpa production systems. Diversity of chilacayote is under study from the SNP and Chalco-Morelos. Collections were made from milpa production systems (corriente type) in 16 localities from eight municipalities (five in Sierra Norte de Puebla and three in Chalco-Morelos) and the fino type was collected from seven localities representing two municipalities in the Chalco-Morelos region. All accessions originated from altitudes ranging from 1600 to 2500 m, but differences between the types were observed in elevational distribution. All of the 38 fino type accessions were collected from elevations over 2200 m. In contrast the 35 corriente type accessions originated from a much wider range of altitudes. Plant characteristics observed to differ notably between chilacayote fino and corriente types include the following: earliness (93 vs. 124 days on average); number of immature fruits (923 vs. 102 fruits per plant on average); number of mature fruit (130 vs. 17 fruits per plant on average); yield of immature fruit (173 vs. 30 kg per plant); and yield of mature fruit (130 vs. 115 kg per plant). Two regional corriente types (referred to here as TeTe and ZaCo) and three fino types were identified. The extreme earliness and high yield of fruit in corriente type C. ficifolia landraces was previously unreported. Within accession variability was observed, especially with regard to earliness. On average the fino types flowered earlier and were higher yielding, but they also tended exhibit less variability than the corriente types. Diversity of Cucurbita pepo, C. moschata, and C. argyrosperma Biodiversity analysis to date has focused on three approaches. First, as a complement to the farm-based breeding research, more than 200 landraces have been collected and deposited in the UACH germplasm bank, and then characterized and evaluated under experiment station conditions. Second, detailed analysis of intrapopulation genetic diversity is being conducted on a set of three landraces, one representing each of three different species from different localities: C. pepo Chapingo (Mexico state); C. moschata Jantetelco (Morelos); and C. argyrosperma Achichipico (Morelos). For the latter three landrace populations, genetic variances and their components are being estimated through half-sib family evaluation. Third, genetic variation and its components at an interpopulation level is being studied through a synthetic population formed by landraces of C. pepo that originated from both of the target regions. The method used for collection consisted of targeting three farmers per location with a total of three fruits representing each Cucurbita species per farmer. The characterization at UACH utilized 40 IPGRI descriptors and additional evaluation criteria (mainly incidence and severity of diseases). Preliminary results revealed high phenotypic variation within and between landraces for the traits of interest. Mature fruit quality characters differed between MILPA Page 13 species, probably largely as a result of domestication and selection criteria applied over time by farmers. Substantial variation was noted in C. pepo for flavor of fruit flesh, which ranged from insipid to very sweet. In C. argyrosperma distasteful colors and flavors of fruit flesh were predominant probably because the main consumption product of this squash species is seed. Fruit flesh in C. moschata tended not to have a desirable color (38%). The latter species exhibited variability for flavor, but was more likely to be insipid than were the C. pepo landraces. Several of the characteristics of fruit and seed yield and fruit quality studied in the landraces showed great variability within and between varieties. Notable is the variation in seed weight per fruit, color and flavor of pulp, and fruit weight. The large sample size (up to N > 200 fruits) required to study those characters is an indirect indication of the magnitude of variation, especially for seed yield and fruit weight which are among the most complex characters. Results of a disease-resistance evaluation trial that included squash landraces from the MILPA target regions indicated great variability within and between Cucurbita species in reaction to virus and fungal diseases, measured in terms of incidence and severity. Crop Biodiversity: Quelites Quelites, or edible herbs derived from the milpa, have been considered a source of food in the local farmers‘ diet by subsistence agriculturalists. However, to most outsiders they appear to be weeds because of the invasive nature and apparent competition with the principal cultivated crops. These plants not only contribute to the meals of the farmers but also protect the milpa system (e.g., soil conservation). It is anticipated at the end of the project that we will be able to correlate certain factors of the quelites (e.g., richness, cycle, etc.) with the diversity of the milpa (e.g., productivity, crop diversity) and with the well being of the farmers (e.g., nutritional contribution to the diet). Some of the most common quelite species include Xanthosoma robustum (mafafa), Amaranthus hypochondriacus (quintonil rojo), Rumex crispus (lengua de vaca), Cyclanthera langaei (macuilquilit), and Brassica rapa (nabo). The first phase (years 1-3) of MILPA has focused on the descriptive component of the quelites in the study areas, especially the SNP where they continue to play an important role in the household meals and the market. Preliminary studies in the Chalco region revealed the abandonment of quelites; probably due to herbicide use, but possibly also due to changing food habits may have contributed to disinterest in quelites. During the second phase (years 4-6), more attention will be placed on describing the situation in Chalco region and re-introducing the utilization of quelites in selected communities. The first phase of the quelites program in the SNP has focused on the taxonomic identification, the quantification of taxonmic diversity and biomass, and the response of nondomesticated plants in the milpa to management practices. Each year, different plots were set up and periodically (1 to 2 times per month) monitored and harvested. Two plots systems were used. One system centered on the traditional agricultural practices with four replications. Within each subplot, 4 combinations of agricultural practices were applied: P1 - soil movement (labra) and hilling (aterradura), P2 - soil movement and no hilling, P3 - no soil movement and hilling and P4 - no soil movement and no hilling. The second system had 30 randomly numbered subplots in which 3 one-meter-square sections were harvested MILPA Page 14 each month over the 10 month agricultural period. After each monthly harvest, all the plants were brought back to the lab where the following data were obtained: number of species, number of individuals by species, biomass contribution per species, average total, aerial and root biomass for each species (based upon 10 randomly selected plants), and average aerial height per species. Samples of different quelites were obtained for proximal nutritional analysis by the Instituto Nacional de Nutrición (INNSZ; Mexico City). For each sample, 500 gm (dry) is required. Each sample must represent a single species from the same community grown under the same conditions (e.g., habitat, phenology, time of the year). The initial samples will be those of the crude plant during its phase of edibility and its phase of nonedibility. A sample questionnaire with basic socio-economic data and basic questions on quelites (e.g., types of quelites recognized and used, source areas of quelites, preparations, management systems, etc.) was designed. Information programs are needed in the communities and to educate the public. At the end of the three agricultural cycles (1998), the data from each plot will be compared and a standard sampling will be selected for application in the second phase. During 1997-98 period, the objectives are: 1. Complete the testing of different plots to evaluate the ecological role of noncultivated plants in the milpa system, especially those used as food. 2. Perform a trial application of questionnaires. 3. Initiate the nutritional analysis of selected quelites. 4. Promote education about quelites through community participation (e.g., establish school contacts in the SNP, bring quelites to the attention of the public nationally). The modified questionnaire was applied in the community of Naupan (SNP) where the generalized questionnaire of the Socio-economic Subgroup was made last year. Using open interviews, 15 households have been sampled and the remaining will be completed to bring the number to 44 homes interviewed about quelites. Discussions were held and preliminary agreements were made in the SNP to carry out educational programs on quelites and milpa diversity with the students at the Instituto Nacional Indeginista (INI) boarding schools. About 16 schools are available but only three are being considered due to limitations in resources. Rosaura Chaparro (director) has agreed to work with us starting with preliminary programs in quelites along with maize. Two public programs to sensitize the public were carried out. A national radio program called ―Buen Provecho‖ of UNAM‘s Programa Universitario de Alimentación was broadcast as a live discussion talk show on national radio (Nov. 1997). During the one week ―Festival del Centro Historico, Cuidad de México‖ of March 1998, we will have demonstrations and general publications for the public who visit this annual event. Information derived from this project is being used in the undergraduate theses of Delia Castro, Emiliano Robles, and Roberto Alvarado and the master thesis of Virginia Evangelista. MILPA Page 15 Gene Flow Gene flow refers to the transfer of genes from one population to another or from one species to another. It can occur through natural events of hybridization by transfer of pollen by insect pollinators, wind, or other means. It can also occur through deliberate hybridizations between and within species. Thus gene flow is the essence of plant breeding, both traditional and molecular-based.Gene flow has important consequences for genetic variability within species and for crop improvement. In MILPA both aspects of gene flow are being addressed. Genes from teosinte, the progenitor of modern maize, are being deliberately transferred by backcrossing into maize. Natural hybridization between bean and squash progenitors and domesticated landraces or varieties is known to occur, but how frequently and under what conditions is not known. These aspects are under investigation in MILPA . Gene Flow: Teosinte to Maize Evaluation by Chromosome Knobs of Introgression of Zea diploperennis to Maize Introgression of Zea diploperennis to maize has been accomplished by backcrossing. Evidence that the teosinte germplasm is present in the backcross-derived maize has been obtained by examination of the chromosomes of the backcross derivatives at meosis. Teosinte chromosomes have distinctive knobs that can be directly observed in preparations of pollen mother cells. Results from six BC3 maize-teosinte inbred lines showed that knobs typical of Zea diploperennis are maintained up to the third generation of backcrossing in frequencies from 0.4% for 3L3 position to 16.3% for 1S3 position of 264 chromosomes observed. Assessment of Teosinte as Source for Maize Improvement in Inbred lines. Potential maize improvement by teosinte utilization was assessed through sequential backcrosses to maize. Six inbred lines and three backcrosses to maize were evaluated in replicated experiments at Celaya, Gto.(1996 and 1997) and Tlajomulco, Jal. (1996). Averaging over experiments and inbreds (Table 1), Zea mays ssp. parviglumis (race Balsas from Jalisco) was the highest yielding of all teosinte sources; Zea diploperennis had the lowest yield and the highest values for days to flowering, plant height, and ear height. BC2 (12.5% of teosinte) was the backcross yielding 37% more than the original inbreds; number of ears per plant increased almost three times for the inbreds with 25% of teosinte (BC 1); it increased 65 and 23% in BC2 and BC3 respectively. Days to flowering decreased with teosinte introgression and plant height was increased. MILPA Page 16 Table 1. Means over environments and reps. Maize-teosinte backcrosses. Celaya, Gto. 1996-1997 and Tlajomulco, Jal. 1996. Teosinte source JALISCO OAXACA MAZATLAN CHALCO M. CENTRAL Z. diploperennis DMS Teosinte (%) BC2 (12.50%) BC1 (25.00%) BC3 (6.25%) ORIG. (0 %) DMS N 288 288 288 288 N 192 192 192 192 192 192 YIELD kg/ha 4483 4234 4197 4123 4041 3864 212 YIELD kg/ha 4596 4461 4225 3347 133 EARS/ PLANT 1.69 1.96 2.12 1.71 1.77 1.63 0.12 EARS/ PLANT 1.75 3.15 1.30 1.06 0.09 ROOT LOD(%) 5.45 6.98 6.19 8.08 5.21 6.58 1.30 ROOT LOD(%) 6.05 8.78 5.66 5.17 1.27 STALK LOD(%) 4.43 4.62 4.65 3.62 4.19 5.78 1.32 STALK LOD(%) 4.51 5.09 3.90 4.69 0.83 POLLEN days 79.82 80.38 79.93 77.90 78.58 80.80 0.46 POLLEN days 79.15 78.62 79.40 81.10 0.31 SILK days 81.13 81.41 81.15 79.05 79.96 82.33 0.46 SILK days 80.24 79.22 80.82 83.07 0.33 PL. HT. cm. 195.25 196.15 191.29 192.14 189.13 181.32 3.67 PL. HT. cm. 194.96 209.69 188.22 170.66 1.87 EAR HT. cm. 87.33 88.16 85.09 84.98 83.00 78.72 2.65 EAR HT. cm. 85.15 99.17 82.03 71.84 1.52 The above results for inbred lines show that teosinte germplasm can be introduced into productive maize inbred lines. A significant level of heterosis is needed if these lines are to be used for hybrid maize. A complete set of crosses (Factorial Design II) among various teosinte BC3 lines were tested in 1997 for one heterotic combination (LPC1 x LPC18). Complete sets of BC3 lines were substituted in one side of the pedigree of the outstanding hybrids H-357 and H-358 to test the potential contribution of genes from teosinte for the improvement of maize. Further evaluation of the BC3 teosinte-maize substitution lines in appropriate control crosses will continue in 1998 and 1999. All this work has been done with inbred lines adapted to subtropical environments in Mexico. The crosses for the combination LPC1 x LPC18 were evaluated under rainfed conditions at two locations, Tlajomulco and Tepatitlan, Jalisco. The experimental design in each location was a 7 x 8 rectangular lattice with two replicates; seven hybrids were included as checks. Although the evaluation was based on data of only two locations and two reps. per location (data not shown), some important points that were evident are: Some crosses with teosinte germplasm had means similar to the best check hybrid, H-315. Some crosses with teosinte germplasm in one or both sides of the pedigree out-yielded the control LPC1 x LPC18 by about 10%. Most of the agronomic characteristics of the crosses with teosinte are similar to that of LPC1 x LPC18. These are extremely important results, indicating for the first time the potential value of teosinte in maize improvement MILPA Page 17 Gene Flow: Between Wild and Domesticated Beans Studies of gene flow are faced with an inherent problem of scale. Detailed studies to determine the level of outcrossing (and the factors influencing it) in a particular location or year will generally not give a reliable estimate of the average amount of outcrossing. Conversely, studies of wild and domesticated germplasm of a given region may be able to determine an overall level of outcrossing, but will not be able to assess the variance of the phenomenon across locations and over the years. Mexico Region Seventy genotypes of wild and cultivated common bean were studied: 15 domesticated, 35 wild and 7 weedy from Mexico, 3 wild from Guatemala, 1 wild from Costa Rica, 1 wild from Argentina, and 8 Andean cultivars. DNAs extracted from all these accessions were analyzed with AFLPs resulting in 145 polymorphic markers. The Nei distances between genotypes were calculated and utilized in a UPGMA cluster analysis using NTSYS. In agreement with previous studies, all Andean genotypes clustered separately including a domesticated genotype collected in Mexico, 17 CMx Bulk Cacahuate, which was introduced into Mexico from South America. Moreover, the structure of the cluster of Andean varieties is in agreement with the pedigree information. Within the Mesoamerican cluster, it was possible to identify three groups that were separated at comparable levels of similarity. The first included 18 wild types, all the weedy types, and the domesticated types. The second and the third clusters included only wild types. The second was formed of 15 wild genotypes from Mexico, one from Guatemala and one from Costa Rica. The third included only three genotypes, two from Guatemala and one from Chiapas. Within each of the last two clusters, genotypes tended to be clustered in relation to their geographic origin. The clustering of wild and domesticated accessions from the state of Puebla suggested some level of outcrossing on a local (sympatric) level. Therefore a second experiment was conducted with accessions that included wild and domesticated types growing at different distances from each other. Chiapas Region To obtain plant materials for this analysis, collections were made in December 1996 in the State of Chiapas. Nine domesticated and 12 wild populations were collected on single-plant basis in three different areas. Also collected were two weedy populations from the States of Oaxaca and Durango; four wild populations from the States of Puebla, Morelos, Mexico, and Jalisco. A total of 29 populations (343 different individuals) were studied using 79 AFLP markers. Results were analyzed using Nei‘s (1973) genetic distance between populations and the distance matrix obtained was used for a UPGMA cluster analysis using NTSYS. The UPGMA analysis showed two clusters. The first included all the domesticated populations, two weedy accessions from Oaxaca and Durango, the weedy entries from Puebla, and a wild accession of Jalisco. This cluster was further subdivided in two groups, one with all the domesticated entries from Chiapas with the exception of the determinate-type population (ChiapasTe-matitaC), which is much earlier than all the other entries; the second with the wild or weedy and domesticated populations of Jalisco and Puebla, the weedy populations of Durango and Oaxaca, and the population ChiapasTe-matitaC. MILPA Page 18 The second cluster included the wild populations of Morelos, Estado de Mexico, ChiapasTx, ChiapasTe, and Chiapas-LR , which clustered in agreement with their geographic origin. With the exception of the populations of Puebla, which showed a high level of similarity between wild and domesticated, and to lesser extent the ones from Jalisco, the UPGMA analysis showed no relationship between the level of sympatry between wild and domesticated forms and the respective genetic distances. Field Plantings in SNP Field sites had been set up in Nauzontla and Huapalecan (SNP) in previous years. During 1997, four domesticated bean landrace components originating in SNP were planted as part of the milpa on one of the sites. The four components had different seed colors: shiny black, opaque black, red, and beige. These were surrounded by a weedy bean (local nahuatl name ―pixnet‖) which grows spontaneously in the milpa, as well as two related species, Phaseolus coccineus and P. polyanthus. Observations were taken on activity and flight paths of insect pollinators (Bombus medius, B. weisi, and Xylocopa tabaniformis). These species visit the three local Phaseolus species. P. coccineus, P. polyanthus, and P. vulgaris ―Pixnet‖, which suggests that there may be locally a substantial potential for outcrossing. Concurrently, measurements were taken on the sugar content and volume of nectar in the various materials: 1) P. vulgaris ―Pixnet‖ = 54% sugar, 2.1 l; 2) domesticated P. vulgaris cultivado = 53% sugar, 1.9 l; 3) cultivated P. polyanthus = 44% sugar; 1.8 l; 4) domesticated P. coccineus = 23% sugar, 2.4 l. These surprising results show that there is an inverse relationship with the known ability to cross-pollinate and the sugar content in nectar, since as P. coccineus, an allogamous species shows the lowest sugar levels among the species studied. Scanning electron microscopy studies show that pollen is deposited on the frontal part of the bee‘s head. Serial flower visits may therefore create opportunities for multiple paternity during outcrossing because of mixing of pollen on the insect vector‘s head. To determine the actual levels of outcrossing, seeds and pods were harvested from the landraces and wild beans for analysis at UCD. Preliminary experiments were conducted to identify suitable molecular markers (sufficient polymorphism; nonradioactive detection). ISSRs (inter simple sequence repeats) appear to fit the bill. Markers have been identified that distinguish wild and domesticated beans of the region. In turn, these markers are now being analyzed in the progenies from the field experiments in SNP. Conclusions About Gene Flow in Phaseolus 1. The fact that the structure of the cluster of the Andean varieties matches the available pedigree information confirms the usefulness of AFLPs in genetic diversity studies and their power to show genetic relatedness in common bean. 2. On the basis on the results obtained until now, gene flow between wild and cultivated is an event that occurs rarely, consistent with a low outcrossing rate and with divergent selection between natural and agricultural environments. 3. However, in some cases such as the Puebla populations, gene flow seems to be an important phenomenon, and a study of the level of outcrossing and the factors (ecological, human, etc.) involved is being conducted. 4. From the analysis, it appears that gene flow between wild and cultivated beans is main- MILPA Page 19 ly uni-directional from cultivated to wild rather than bi-directional. In fact in both experiments, it was possible to find a few wild or weedy accessions clustering with the domesticated accessions but never the opposite. Selection conducted by the farmers like the one for larger seeds could explain this asymmetry. 5. Another possible explanation for the presence of wild populations that are closely related to the domesticated forms is that they originated from domesticated materials escaped from cultivation and established in disturbed environments. This can be a possible explanation for the weedy form found in Oaxaca and Durango, which show morphological features of the domesticated type (i.e., seed size and seed color), but it seems more unlikely for the populations of Puebla and Jalisco, which are morphologically similar to the other wild material. 6. According to our results, gene escape from transgenic plants would not seem to be a major and geographically widespread risk, at least if the transgenes or genes linked to the site of insertion do not confer a significant selective advantage in natural environments. Gene Flow: Cucurbita argyrosperma and C. moschata The objective of the gene flow-related research is to identify key factors controlling gene flow within the Cucurbita component of the milpa and to characterize the magnitude of gene flow. Field studies were conducted in the fall of 1996 in the Autlán Valley in Jalisco in two localities (San Lorenzo and Los Parajitos) in milpas belonging to two farmer-cooperators. Two closely related and highly cross-compatible domesticated squash species (Cucurbita moschata and C. argyrosperma ssp. argyrosperma) were present in the milpas as were two weedy taxa: the wild type C. argyrosperma ssp. sororia and putative inter-subspecific hybrids of C. argyrosperma. The latter either occurred spontaneously or, at one site, were also being planted by the farmer as part of his own breeding effort to enhance certain seed traits. The molecular analyses of Cucurbita focused on screening for molecular markers to detect gene flow within populations of landraces and crop wild relatives. So far DNA has been extracted from nearly 25% of the more than 200 samples collected from individual Cucurbita plants in the San Lorenzo and Los Parajitos milpas in 1996. A total of 58 RAPD primers were screened initially using a subset of individuals that represented each taxon and site combination. Eleven of those primers exhibited polymorphisms considered suitable for further analyses. For each of the selected primers from 4 to 12 bands or ―loci‖ were judged to be scorable, ranging from an average of 5.1 to 7.1 scorable loci per primer per taxon. The average number and percentage of polymorphic loci per primer was lowest in the putative hybrids (2.8 polymorphic loci per primer; 55.4% of all loci for that taxon) and in C. argyrosperma ssp. argyrosperma (3.0; 49.3%), slightly higher for C. argyrosperma ssp. sororia (3.5; 62.5%), and highest in C. moschata (4.7; 67.1%). Shannon-Weaver indices depicted a similar pattern: lowest for domesticated C. argyrosperma (0.36) and the putative hybrids (0.37), intermediate for the C. argyrosperma wild type (0.45) and highest for C. moschata (0.51). RAPD primers look promising to detect C. moschata markers, but not for C. argyrosperma subspecies. RAPD primers tested did not reveal any markers to aid in distinguishing wild from domesticated C. argyrosperma germplasm. Genetic identity values (Nei 1972) were 0.97-0.98 among the C. argyrosperma subspecies and the putative hybr- MILPA Page 20 ids in striking contrast to values of 0.67-0.68 when comparing those taxa to C. moschata. The close genetic similarity between the subspecies of C. argyrosperma as revealed by the RAPD primers is intriguing in terms of potential ramifications for farmers attempting to produce their squash crop in an environment surrounded by a barely edible yet highly cross-compatible wild relative. During 1998 a larger set of samples from the Jalisco populations will be analyzed for the RAPD primers. Additionally, a molecular study using ISSRs has been initiated to determine if any of these genetic markers might be found to distinguish the wild and domesticated counterparts. The pollination system in Cucurbita is being investigated by observing insect pollinators and their behavior, Cucurbita flowering phenology, and experimental and natural hybridization among the Cucurbita growing in the local milpas. An undergraduate thesis on the pollinator studies was completed in 1997 (A. Bautista-Anaya, Univ. Guadalajara, Autlán). In two milpas the total number of censused bee visits and the dominant bee species were recorded. A Ph.D. study will be initiated by Montes in the coming year to further explore gene-flow dynamics in milpa-grown Cucurbita, using the Jalisco milpas as an initial base. MILPA Page 21 Objective 3. Develop and evaluate on-farm breeding methods to improve the productivity of local landrace germplasm through mass selection or introgression from improved germplasm (maize) or wild or cultivated relatives (beans, squash). On-Farm Mass Selection in Maize Many breeders have demonstrated that simple mass selection within landraces of maize can result in genetic improvement in grain yields. Mass selection is defined by the fact that the selection unit is a single plant. The key to achieving gains with this form of selection is to reduce environmental noise to maximize the probability that superior individuals are superior because of their genetics rather than because of favorable environment. Mexican maize farmers typically select their seed for planting from among ears amassed on a drying floor. The best ears are chosen and pooled to make the seed supply for the next season. This technique has been applied for centuries and no doubt has resulted in improvement, albeit slowly and of modest gain, in maize performance. Since this is such an old practice, future gains are expected to be small, and the selection procedure assures a stabilized population will not decline in its genetic potential. The MILPA project is evaluating an intervention to the farmers‘ age-old practice by initiating selection in the field before harvest. This includes selection of healthy plants and then selecting the best plants and ears from those plants from a defined area in the farmers field. The farmer‘s field is partitioned into sections and each section becomes the unit where selection is done. The defined section contains a certain number of hills, about 100. The reason for using the grid system as a unit for selecting the best single plants or ears is that all of the plants in a grid have grown under rather uniform conditions so that differences among them is more likely due to genetic effects than environmental effects. Thus, the farmers make their initial selection of ears in the field, taking the best ones from within each section or grid. These can be pooled to form the seed lot for the next planting. For MILPA, this scheme is being tested on farms with repeated selection planned for 4 or 5 years (generations). The accumulated gain from such selection will be demonstrated by growing plants in a common field from each of the previous selections along with the farmer‘s local seed source. If significant gains are demonstrated, this practice will be taught and offered to all farmers in the region and to others throughout Mexico. Developing the method necessarily is done on farmers‘ fields. The amount of involvement of the farmers in the experimental phase may vary depending upon their own interests in participating. Surely, a special interest is taken to determine their attitude about in-field selection and that can be judged during the experimental phase and lessons can be learned from the farmers about how to carry out this intervention with minimal disruption in their current farming practice. Progress has been made during the past three crop seasons. Five populations have been mass-selected in the Chalco region using three of the recognized subtypes of Chalqueño at farms at different altitudes. Similarly, in SNP three landraces in two communities are under mass selection. Already, the attitudes of farmers about participation in this process has been noted. Some of them are enthusiastic and are already involved in the selection process and feel that gains have already been obtained; others allow that the selection experiments be done on their farms, but are not confident about the expected results. Preliminary results have been obtained from comparisons after one cyle of mass selection MILPA Page 22 was completed in the Chalco region. Based on controlled experiments done over 10 to 20 generations by breeders in the US and elsewhere, the gain realized from each generation of selection is expected to be small, and probably difficult to detect without having benefit of seed from several cycles included in the same trial. A selection response curve can be tested statistically and an average gain per generation can be computed. The one-cycle test done in the Chalco in 1996 provides encouragement but not proof that ‗farmer-significant‘ gains will be realized. Grain yield from Cycle 1 was 2792 kg/ha compared to 2717 and 2641 for the original population from 1995 and 1994 seed source. The ‗gain‘ of Cycle 1 over the original was 5.7 and 2.7% for the two comparisons. This gain was not statistically significant, but within the range expected. The Cycle 1 plants showed a desirable 3% reduction in plant height, lower ear height, and about 1 day earlier to silking and tasseling. Trials are in progress to continue the assessment of progress made by mass selection, but the first results are encouraging. Introgression of Genetic Diversity from Wild to Domesticated Beans So far, F2 and F3 populations from biparental crosses have been developed, as well as BC1F2 and BC2F1 populations that are being advanced and multiplied for testing in the field during 1998. During the winter of 1997-98, two nurseries were established at Zacatepec, Morelos. One nursery included 35 F2 simple crosses, cultivated x wild, or wild x cultivated (most of the biparental crosses were made in both directions, i.e. with the cultivars as female and male parent). The other nursery included 24 BC 1F2, 15 F2 double recombinants (F1 x F1 of the same cross), two four-way crosses, and four F3s derived from biparental crosses. Most of the populations will be mass-harvested and in a few of them individual plants will be extracted at random for future studies. Our observations identify a difference in performance between crosses with wild or weedy bean accessions. It remains to be determined from which type of accessions, wild or weedy, introgression will be achieved most effectively. The yield data show that evaluations in F2 and later inbreeding generations are inadequate to evaluate the potential of wild x cultivated crosses. Instead a backcrossing program to the cultivated parent has to be conducted first to eliminate a majority of the deleterious traits from the wild parent. For long-term research, utilization of the heterotic effect in beans seems to be a major opportunity for yield advances. A preliminary yield trial of 11 F2 populations derived from biparental crosses between cultivated x wild P. vulgaris was conducted under rainfed conditions with supplemental irrigation in the summer of 1997 at Texcoco, State of Mexico. Three cultivated parents: Negro Tacaná, Puebla 152, and Bayo Baranda and the elite cultivar Bayo Mecentral were included in the trial. Plant stand was similar to commercial planting, i.e. 120,000 plants/ha. In addition to days to flowering and reaction to naturally occurring diseases, seed yield of individual 2 plants and per m were measured. Germination was slow in the segregating populations as compared to cultivars and the final plant stand varied from 50 to 80% in these populations. Since none of the seeds were scarified, the differential germination within populations might be due to shell impermeability or intrinsic dormancy as displayed in wild stands. In some of the populations, few intact seeds and young seedlings were still observed at 80 days after planting. Although no record on biomass accumulation was made, it was noticeable that the populations derived of MILPA Page 23 weedy types were more vigorous than those derived from wild types. Cultivars showed superior seed yield when compared to the segregating populations. Among the segregating populations, the top yielder was the one derived from Negro Tacana x weedy from Hidalgo (W10). The low yield of segregating populations was due to the display of many dominant traits derived from the wild parent, such as lateness, aggressive vegetative growth, and small seed size. Squash Breeding Maize herbicides are mentioned by the local farmers as the primary explanation for the reduction of squash production area because squash is susceptible to herbicides. The result is that squash is being eliminated as a component in the milpa cropping system. Díaz (1997 UACH undergraduate thesis; McKnight MILPA Project) found variation for herbicide resistance to atrazine and linuron, two typical maize selective herbicides, in a C. pepo synthetic line obtained by genetic recombination of landraces from different geographical origins. This discovery suggests the possibility for improvement of squash via selection for resistance to maize herbicide, thereby retaining one of the important intercropping components of the milpa. The MILPA squash improvement component involves: (i) on-farm selection of local landraces in a series of projects initiated in 1996 and conducted in association with cooperating agriculturists using selection criteria derived from joint decision by local farmer-cooperators and the breeder and (ii) introduction of desirable traits into landraces of three species (C. pepo, C. moschata and C. argyrosperma) through hybridization with improved germplasm. Both types of breeding are aimed mainly at selection for improved fruit quality and fruit and seed yield. Intrapopulation Selection Based on field collection activities, eight cooperating farmers were selected to initiate participatory plant breeding. Due to the nature of the study and sampling requirements, it was necessary that each farmer-cooperator have at least 2 ha of squash under production in association with maize. Four farmers with C. pepo were selected in the Chalco region and two others with the same species in the SNP. A single farmer with C. argyrosperma and another with C. moschata were selected in the hotter, lower elevation area of Morelos. The breeding work entailed a selection pressure of 5% with seed and fruit yield and fruit quality traits as the selection criteria. Although in 1996 mass selection was applied in certain fields (i.e., those in Chalco-Morelos), for most sites the true breeding work began in 1997, so the results reported herein are preliminary. These landraces improved by participatory plant breeding will be available to farmers in the last phase of the MILPA Project. After this first cycle, selection responses depended on the trait under selection and the locality. For C. pepo, the selection differentials for fruit weight ranged from 310 to 763 g per fruit, whereas seed yield ranged from 6.7 to 33.8 g. In one landrace there was a relatively large reduction of weight per fruit. This occurred because the farmer initiated a change in the selection criterion from fruit size and good fruit quality in 1996 to higher seed yield in 1997. In C. pepo landraces from SNP (El Tejocote and El Socorro) only yield of fruit and seed weights were determined. The responses to selection for fruit and seed yield in C. argyrosperma were 180 g per fruit and 94.0 g of seeds per fruit. In C. moschata the corresponding gain was 395 g and 77.2 g. For the latter landrace population, a gain in 0.73 cm MILPA Page 24 was noted for pulp thickness. These selection responses were remarkably large and indicative of considerable genetic variability within the landraces and that the traits under selection had high heritability. Additional cycles of selection are in progress. Interpopulation Breeding Through crosses and partial backcrossing, desirable agronomic characters such as earliness, bush type, and hull-less seed are being introduced to landraces. The source for earliness and bush type is C. pepo ‗Round Zucchini‘. In 1997 F2 seeds were obtained from crosses of local germplasm x hull-less seeded C. pepo crosses. The landraces used in the crosses consisted of 26 C. pepo, 7 C. argyrosperma, and 1 C. moschata. An additional 20 crosses (F2 seed) of C. pepo landraces and C. argyrosperma x ‗Round Zucchini‘ were obtained. A cross from the pumpkin cultivar ‗Jack Be Little‘ (C. pepo) by a C. pepo landrace resulted in high heterosis for yield and quality (good texture, orange color, and thick fruit flesh) of fruit and size in plants characterized by extreme earliness. This line and others possessing the bush habit, earliness, hull-less seededness, high yield, and good quality will be ready in following years for on-farm evaluations by the farmers themselves. IV. Unique role McKnight project has played in doing new research. The 1996 MILPA Annual Report acknowledged the following areas by which this program has uniquely facilitated new research relevant to improving the understanding of crop diversity and human decision-making factors as we endeavor to enhance genetic resources conservation and increase yields, stability, and income for the agricultural sector and thereby enhancing the whole economy of Mexico. These points are still valid for 1997. A more comprehensive study of gene flow between crops and their wild relatives, with emphasis on maize, bean, and squash. This topic is of particular importance for genetic resources conservation but also for to evaluate the potential of gene escape from transgenic crops. The development of more comprehensive models to understand household management of crop diversity. This model includes the role of interhousehold linkages as well as household decision making and consumption as well as production factors. A gradual integration of biological and socio-economic research programs and scientists. The role of wild relatives in improving the yield potential of crops, which is usually considered not to be novel enough by federal research agencies, although it holds considerable promise. Alternatives for farmer participation in improvement of local crop populations. V. Training During 1997 several new students joined the project, and advanced training in the US was supported for two Investigators. At present there are 12 M.S. students in Mexico and 9 Ph.D. students (four in Mexico and five in the US) (Table 2) and 16 undergraduate students in Mexico and one postdoc in the US supported by MILPA. Cristina Mapes finished her Ph.D. at UNAM and Cristina Ugarte finished her M.S. at UNAM. Advanced training for investigators MILPA Page 25 Antonio Yunez Naude from the Colegio de Mexico, Sabbatical leave with E. Taylor, UC Davis, 1996-97. Alfonso Delgado from UNAM, Sabbatical leave, Montana State University, 1997-98. Postdoctoral fellow Alonso González Mejía, Department of Agronomy and Range Science, UCD, 1997-98. Table 2. Student training in degree programs. Name Ph.D. Students Edgar B. Herrera J. Francisco Casas Roberto Miranda Hugo Perales Margarita Mauro Eric Van Dusen George Dyer Salvador Montes Miguel Sandistevan M.S. Students Cristina Ugarte Ileana M. Nuñez Valdemar Ballesteros Alfredo Wong Miguel A. Sánchez Isaac Meneses Francisco Basurto Xochitl Juarez Roxana Martínez Javier Becerril Virginia Evangelista a b Instit. CP CP UG UCD UCD UCD UCD UNAM UCD FIU CP CP UNAM UACH UACH UNAM UACH CM CM UNAM Year 1998 1998 1998 1998 1999 1999 2000 2001 2001 1997 1998 1998 1998 1998 1998 1998 1998 1998 1998 1998 a Project maize biodiversity maize / teosinte gene flow & breeding maize / teosinte ethnobotany socio-economics (maize) bean breeding (molecular markers) socio-economics (milpa) socio-economics (milpa) squash gene flow socio-economics (maize) quelite ethnobotany maize biodiversity maize biodiversity bean gene flow squash biodiversity & breeding squash biodiversity & breeding quelite ethnobotany socio-economics (maize) socio-economics (maize) socio-economics (maize) socio-economics (maize) Locality b Chalco Jalisco Jalisco Chalco, Morelos SNP, other SNP SNP Jalisco, Morelos Chalco, other SNP Chalco Chalco SNP Morelos Chalco SNP SNP SNP SNP SNP Year of completion for degree program (actual or expected). Source of germplasm or locality for on-farm research. VI. Additional support The following remarks, made in the 1996 MILPA Annual Report, were relevant in 1997. It was anticipated that graduate students would be supported by CONACYT and would join this project. In spite of reduction in CONACYT sponsorship in Mexico, there are seven students who have received funding from CONACYT or another source in Mexico. The CONACYT support per student is about $7,000 per year. For the students studying in the US, their stipends are supplemented by MILPA funds. The universities in Mexico are also providing support for graduate students that do not have CONACYT fellowships. As the project is advanced to the stage that outreach education programs will be appropriate to diffuse methodologies, such as on-farm plant improvement, we will solicit assistance from Mexican extension service and from NGOs and foundations for implementation of these activities. MILPA Page 26 VII. Advisory committee activities The project participants are appreciative of the efforts and dedication of the project by those of the AC who have been able to participate. AC suggestions have been valuable and heeded during the execution of the project and in this renewal proposal. This is reflected in the emphases on coordinated and integrated activities by different crop and disciplinary specialists and on implementation and evaluation. The AC emphasized the need of the project to move beyond the first, descriptive, phase of the project and to initiate more active efforts in participatory crop improvement. This message was very influential in the planning and design of the renewal of MILPA. VIII. Limitations and constraints Several constraints have become apparent over the course of the project. The milpa cropping system may have become a somewhat idealized concept, in that a push for monoculture in various parts of the country, principally towards maize, has made it increasingly difficult to find common sites for the various subprojects. This is even the case in the Sierra Norte de Puebla, a more traditional area. Timely communications among MILPA participants has been limited and it is clear that this activity must be enhanced. The internet web site is increasingly used, especially as Mexican colleagues acquire internet access. The collaboration between biological and socio-economic researchers still requires encouragement. Challenging scientific tasks, long traditions of working in isolation from each other, and different research traditions and cultures all affect the amount of interdisciplinary collaboration. Nevertheless, collaboration has increased during the past year. We see this project as an opportunity to learn to work together and to educate students in this respect. MILPA Page 27 IX. Assessment of progress, future plans, revision of goals The Annual meeting in October and the preparation of the project renewal proposal in November allowed us to evaluate the status and progress of this large and complex interdisciplinary project. There was a real sense of accomplishment during these two meetings. First, we recognize that we have achieved a much higher level of interdisciplinary communication and collaboration than seemed possible after Year 1. Second, we realize that we have made significant progress in establishing a baseline from which to assess progress. This is especially true in the Chalco/Cuautla study region, but achieving a baseline is also now within grasp in SNP. Third, substantial progress has been made in the basic descriptive work of assessing biological diversity in the three crops. This is essential to achieve the larger goals of achieving gains in crop improvement whilst maintaining biological diversity. The second three-year phase of MILPA will (1) continue the socio-economic, diversity, and gene-flow research through years 4 and 5, (2) continue the on-farm mass selection for yield and plant type improvement in maize and introduce new landrace-based materials of bean and squash to farmers for selection and evaluation, (3) establish a series of demonstration sites with farmers‘ participation to evaluate the improved products of the MILPA project, and (4) provide an assessment of the importance of quelite species in the diets of rural families and develop outreach activities to promote the value of the quelite species in the milpa farming system. MILPA is clearly contributing to the human resources development of Mexico, while the research is yielding both new methodologies for crop conservation and improvement and new plants for the farmer‘s field in Mexico. These are achievable results within the six-year period of CCRP and MILPA. MILPA Page 28 Appendix I. 1997 Report of the project Advisory Committee (prepared following the 1997 annual meeting October 1997) —Prepared by Peter Bretting, CoChair MILPA Advisory Committee During 1995-97, the AC reviewed the initial MP project proposal, participated in the MP annual meetings, and helped train MP participants. The AC assisted the MP with establishing channels for transferring funds to Mexican institutions. The AC suggested that programmatic PROCESS is at least as important as its products. Consequently, it encouraged the MP to organize its planning, decision-making, and project implementation into an integrated, interdisciplinary team, and the MP has begun to do so. It encouraged cross-subject meetings each year in Mexico, and cross-disciplinary journal clubs for students. It suggested that project time line diagrams with milestones, deadlines, etc. should be established, reviewed and updated by the MP participants regularly as a group. It suggested that research sites be mapped by global positioning systems, that site-specific information be integrated via geographic information systems whenever feasible, and that provisions for the long-term archiving of germplasm collections, scientific data, etc. be established. Also, one AC member (Bretting) helped train an MP participant (Ortega Paczka) in plant genetic resource management and information technology during the latter‘s four-month stay at Iowa State University. Role of the AC in preparing the MP project renewal proposal: During the 1997 Annual Meeting of the MP, the AC focused its efforts on assessing the MP‘s progress toward 1) meeting the McKnight Foundation‘s evaluation criteria for project renewal; 2) attaining the overall project goal and objectives; and 3) incorporating the previous suggestions made by the AC into its operations. The AC recommended that data collection associated with the first two MP objectives (description and analysis of the relationships between farmer knowledge, socio-economic factors and genetic diversity in the milpa agroecosystem; characterization of the structure of crop biodiversity and the magnitude of gene flow from wild or cultivated relatives to maize, bean, and squash) should be essentially completed by the end of year three of the MP, i.e., by March 1998. The AC suggested that, if renewed, the next three years of the MP should focus on goal 3 (development and evaluation of on-farm breeding methods to improve the productivity of local landrace germplasm through mass selection or introgression from improved germplasm [maize] or wild or cultivated relatives [beans, squash]), armed with the data, concepts, ideas, etc., developed during the first three years. The AC also suggested that the next phase of the MP should probably focus on implementing the preceding methods and strategies in two or three sites where most of the MP‘s effort would be centered. Projected roles of the AC in future MP activities: Should the MP be renewed, the AC will continue to review work plans, provide advice regarding overall project direction, attend annual meetings, etc., as during the MP‘s first three years. In particular, the AC may concentrate on helping MP participants form interdisciplinary teams, and on fostering ever stronger inter-institutional partnerships among agricultural research organizations in Mexico. With the help of MP staff, the AC will strive to incorporate farmer-participants as formal AC members, because for the MP to attain its objectives fully, farmer-participants should play an ever more important role in managing and directing MP activities. At least one member of the AC (Bretting) may help train another MP partici- MILPA Page 29 pant (Mapes) in plant genetic resource management and genetic marker technology, should funds be available for Mapes to visit the US.