Providing Farmers the Technology Required to Efficiently Breed Corn Varieties Specifically Adapted to Alternative Cropping Systems

Final Report for LNC03-223

Project Type: Research and Education
Funds awarded in 2003: $134,400.00
Projected End Date: 12/31/2006
Region: North Central
State: Nebraska
Project Coordinator:
W Ken Russell
University of Nebraska
Expand All

Project Information

Summary:

The primary educational objective was to train 20 farmers in modern principles of corn breeding. This was accomplished by two workshops followed by the farmers conducting evaluation trails on their farms. Participation was 90% and 70% for the two workshops and 65% for the evaluation trials. The primary research goal was to evaluate the effectiveness of a new testing protocol that could be implemented by farmers without making significant capital outlays. In terms of data quality, the new protocol was found to be comparable to the standard testing protocol used by seed companies.

Introduction:

The premises of this project were
1) most corn farmers have a poor understanding of the modern principles of corn breeding that professionally trained corn breeders have used to increase by nearly five-fold corn grain yields over the past 75 years, and
2) corn farmers will benefit from an enhanced understanding of these principles.

The first of these was based on the many interactions the project leader has had with farmers during his 20+ years as a private and public corn breeder. The second premise was based on our belief that the private corn seed industry meets the seed needs of most, but not all farmers. Not meeting the seed needs likely is especially true for farmers who are using non-traditional cropping methods. The growth in the organic market is but one example that an increasing number of farmers are turning to non-traditional cropping methods. As these numbers grow, so to will the need of farmers for education in plant breeding.

Professional plant breeding typically is done in high-input environments because the highest ratio of genetic to non-genetic variation occurs at such environments (Bänziger and Cooper, 2001). However, when different varieties are the best in high- and low-input environments, then the top-performing varieties will not be identified for all farmers by selection only in high-input trials. Smith (1991) compared four cropping systems in New York, one high input system and three lower input systems. She found that a single variety of corn was not best for all systems. Bänziger et al. (1997) also reported low correlations between performances of corn varieties in low- and high-input environments.

A second consideration in ranking varieties is the selection criterion. Seed companies focus on harvestable grain yield. One consequence of this intense selection for yield has been a decrease in the protein concentration in the grain (Duvick, 1997). If protein content or other traits of corn not currently being measured by private breeders are important to a farmer, then it is likely that a selection program focused on these criteria will result in better varieties for that farmer than commercial hybrids selected solely for improved grain yield.

Even assuming the hybrids developed by seed companies are the highest performers in all cropping systems used by farmers in the U.S., highest performance does not necessarily translate to best performance based on economic return. Benbrook (2001) estimated that the dollar value added by Bt corn via higher yields from 1996 to 2001 to U.S. farmers was $567 million, but the net premium paid by farmers for Bt corn during this period was $659 million. The net cost to farmers was $92 million. This loss does not include the indirect costs of growing a GMO hybrid, such as costs for segregating different GMO hybrids or a GMO from a non-GMO hybrid.

Seed companies are now beginning to sell corn hybrids in which multiple, genetically-engineered traits are stacked. They claim that these traits lower input costs, such as the cost for pesticides. However, for low-input farmers who already use reduced levels of such inputs, this is less of an advantage.

Additionally, farmers having the scientific knowledge to develop their own corn hybrids may provide some less obvious benefits to American agriculture besides having hybrids that better meet their needs. One likely benefit will be increased genetic diversity. Because most new seed company hybrids and the germplasm from which they are produced have legal protection, farmers will need to use different sources of germplasm to develop their hybrids. The increased genetic diversity that will result from this activity is important because most recently developed commercial hybrids are descended from a base of only six to eight hybrids (Tallury and Goodman, 1999). Secondly, we believe the ability to create improved varieties for their own use will give farmers a feeling of enhanced empowerment that will boost their self-image and morale. Kerr and Kolavalli (1999) recognized empowerment as an important condition for sustainable agricultural communities.

If a farmer has a current or potential need to develop corn varieties, how does he/she proceed? To be successful, he must have an understanding of the same key principles that professional corn breeders have successfully used during the past 80 years. Secondly, a farmer must have a protocol for developing improved varieties that he can implement on his farm without making major capital investments.

Project Objectives:

This project had two primary objectives, one educational and one research-oriented. The educational objective was to train a group of farmers in the modern principles of corn breeding. The most important of these are use of hybrid vigor, family-based selection, replication in field trials, and use of instrumentation to measure small differences in performance among varieties. But the farmer also needs a cost-effective protocol for implementing these procedures. Development and production of single-cross hybrids as done in the seed industry is capital intensive. Therefore, our research objective was to test a protocol a farmer could afford to implement to develop his own varieties.
Meeting these objectives should lead to the following short-term outcomes:
1) An ability by farmers to initiate corn variety development programs utilizing enhanced breeding methods and better field-plot techniques;
2) An enhanced capability of sharing and analyzing farmer-generated data from evaluation trials;
3) An increased awareness of public varieties of corn that may be used as source material in a selection program;
4) A more positive attitude toward the possibilities of farmer-developed corn varieties.

To the extent these outcomes are achieved, we believe the long-term outcome will be the implementation of on-farm breeding programs based on sound breeding principles and ultimately a greater number of corn varieties better suited to sustainable corn production systems and/or that better satisfy a variety of niche markets. This will enhance the economic viability of alternative farming systems.

Research

Materials and methods:

The first step in this project was recruitment of farmer participants. Via regular quarterly mailings that the Nebraska Sustainable Ag Society and also the western Iowa chapter of the Organic Crop Improvement Association send to their members, we distributed over 400 flyers announcing this project and inviting farmers to apply. We also advertised on a University of Nebraska website and at county extension offices throughout Nebraska. The result of this advertising campaign was 33 applications – 19 from Nebraska, 1 from Kansas, 7 from Iowa, I from Illinois, 1 from Indiana, 2 from Ohio, 1 from Maine, and 1 from the Phillipines. This was a lower response rate that what we desired. Also, despite targeting farmers who had interest or active involvement in using organic or non-traditional cropping practices, most of the applicants were growing only single-cross corn hybrids using high-input procedures. We chose 20 participants – 11 from Nebraska, 1 from Kansas, 4 from Iowa, 1 from Illinois, 1 from Indiana, and 2 from Ohio.

We planned to conduct five primary activities with the farmer participants to achieve the objectives:

1) A classroom workshop to introduce key principles in modern corn breeding. This one-day workshop was planned to be held at the University of Nebraska in March, 2004. The topics we intended to cover in this workshop included

–What is plant breeding?
–Early history of corn breeding in the U.S.
–Heterosis (hybrid vigor)
–Types of corn cultivars
–History of corn grain yields in the U.S.
–Phenotype, genotype, environment, and heritability
–Mass and family selection
–Micro- and macro-environmental effects
–Genotype x environment interaction
–Field design and data analysis
–Number of selected families
–Multiple traits
–How the private seed industry conducts corn breeding
–How we propose to breed corn in this project
–Some potential corn source populations

The intent was to cover each of these topics only in sufficient detail so that the farmers would have a general understanding of the science behind the procedures used by professional corn breeders. We organized the topics around class activities in which the farmers would actively participate. For example, we planned for the farmers to estimate heritability by having them determine the variances of ear weight for two inbreds, their single-cross hybrid, and the F2 generation produced from the hybrid and for the farmers to observe the greater variance for the F2 compared to the non-segregating generations. For each farmer we prepared a 63-page hand-out in a sturdy binder that contained notes on each of the topics and also room for them to record the results from the classroom activities.

To assess the success of this workshop, we prepared a 17-question survey (Appendix A-1) that we intended to send to each farmer approximately two weeks after the completion of this workshop.

2) A field-oriented workshop to re-enforce principles introduced in the classroom workshop. This workshop was scheduled for September, 2004 at the University of Nebraska. Much of the farmers’ time at this workshop was to be spent touring a demonstration plot of over 100 entries that was designed to re-enforce many of the concepts that had been introduced during the first workshop (Appendix A-3). As at the first workshop, handouts were prepared that contained a brief description of each entry and also included 18 questions that were designed to test the farmers’ understanding of the key principles in corn breeding. Some of the entries were various types of public corn cultivars for which the farmers could order seed and use to develop new varieties.

Another part of this workshop was a tour of the breeding nursery where seed of the topcross families that the farmers would evaluate the following summer was produced. Although the farmers had no part in producing this seed, our intent was for them to understand how it was produced.

3) Development of electronic lessons to support the learning of the concepts presented in the two workshops. Our goal was to develop lessons that would be housed at the University of Nebraska Library of Croptechnology website. The topics of these lessons would correspond to the topics covered in the first workshop. Each lesson would include a bank of questions to monitor the mastery of the learning objectives. During the second workshop, we planned to take the farmer participants to a computer lab and show them not only how to access this website but also to introduce them to the structure of the lessons.

4) A field evaluation of a large number of topcross families as would occur in an actual breeding program to develop a superior corn variety using a protocol that could be implemented by a farmer. The selection protocol we intended to use was based on the evaluation of 312 topcross families using replicated single-plant plots (henceforth call microplots) that would be hand-planted and hand-harvested. Each farmer participant who indicated a desire to have a test plot on his farm would be provided with a hand-planter and an electronic scale for weighing the ears. An introduction on laying-out, planting, and harvesting a plot was provided at the second workshop and this was followed up with detailed written instructions (Appendices B-1 and B-2). Additional instruction in harvesting the plots was provided during pre-harvest visits to nine of the farms.

The population that was the source of all the topcross families was developed at the University of Nebraska-Lincoln from 2000 to 2002 and has been named NEL. It was developed from six other populations that were identified with good potential for improving the yield of U.S. Corn Belt hybrids (Dudley et al., 1996). Although the genetic make-up of NEL is approximately 60% exotic, it is adapted to the length of the growing season in the central Corn Belt.

The topcross families were produced at the University of Nebraska in 2004 by self-pollinating non-inbred plants of NEL and also crossing each plant onto several plants of an inbred tester (FR1064). Approximately 900 sets (i.e., a self and its associated crosses) were produced, and at harvest the best 312 sets were selected based on the phenotype of the parental NEL plant. The seed produced by crossing each NEL plant to the tester comprised a topcross family and was used in the trials. The seed produced by selfing each NEL plant was a S1 family, and this seed was placed in cold storage. After the best topcross entries were identified, the S1 families associated with these entries were crossed to each other to generate an improved cycle of the NEL population.

The goal was for each farmer to plant a two-replication trial of all 312 topcross entries in the summer of 2005. These were organized into 12 tests of 30 entries each, 26 topcross and 4 check entries. Each check was a single or three-way cross in which FR1064 was one parent and the other parent(s) was a privately developed, elite inbred line. Assuming a plant density of 30,000 plants per acre, the approximate area of a trial was 0.4 acre. An additional three trials were planted by the project leaders at three environments on University of Nebraska land. To compare the efficiency of the microplots to the more traditional or standard plots that are widely used in the seed industry (each plot is two rows x approximately 20 feet in length), three of the tests also were planted as standard plots.

For each of the farmer-planted plots, we intended for the farmers to record data on ear weight (using electronic scales that were provided) and lodging. For microplots planted and harvest by the project leaders, the same data were determined as at the farmer trials and then following drying in a 160 degree Fahrenheit oven for three weeks, dry grain and cob weights also were obtained. From these data we were able to estimate grain yield at 15% moisture. Lodging, grain yield, and grain moisture were determined for all the standard plots.

5) A final workshop during the late winter following the evaluation trials to discuss the analysis of the data gathered in these trials. The goal of this one-day workshop was to demonstrate how basic statistical procedures could be implemented using actual data the farmers had gathered. Our intent was to develop some “cookbook-type” software the farmers could use to undertake these procedures.

6) Other activities. Three additional activities were undertaken that did not directly involve the farmers. First, 3 of the 12 of tests grown in 2005 were re-evaluated in 2006 at three environments as both microplots (six replications per environment) and as standard plots (two replications per environment). A second activity was the production of an improved cycle of the population NEL in the summer of 2006 using the results from the microplot evaluations. The final activity was the development and production of a large number of various types of hybrids during the summers of 2005 and 2006 for testing in 2007. We believed that farmers developing their own varieties should utilize hybrid vigor, but because of cost it is unlikely the production of single-cross hybrids will be feasible for most. Other types of hybrids include population x population crosses, population x inbred crosses, and double-cross hybrids. Using modern germplasm, the advantages and disadvantages of these other hybrid types compared to single-cross hybrids needs to be determined.

Research results and discussion:

(organized by the activities listed under Materials and Methods section)

1) First workshop.
The success of this workshop was measured by two metrics, attendance and the results of the survey. Eighteen of the 20 farmers attended. Both farmers who were unable to attend were contacted after the workshop and both indicated a desire to stay involved.

Twelve of the 18 farmers responded to the survey (Appendix A-2). Overall, the responses indicated a high rating of the workshop. Of particular interest in measuring the success of the workshop in imparting knowledge of corn breeding principles were the answers to questions 5 and 6. The most common rating of the farmers of their knowledge of corn breeding before the workshop was “somewhat low”, whereas after the workshop the most common rating was “average”. Before the workshop, there were two “low” ratings and no “high” ratings, whereas after the workshop there were no “low” ratings and two “high” ratings. Furthermore, the answers to questions 7 and 8 indicated a definite increase in the interest level of the farmers in corn breeding.

2) Second workshop.
Fourteen of the 20 farmers attended. Also attending were two farmers from Nebraska who were not originally members of the project, a corn breeder from Iowa State University, a corn breeder from Michael Fields Institute in Wisconsin, and six plant breeding graduate students from the University of Nebraska.

The success of the demonstration plot as a learning tool was measured by the farmers’ success in answering the questions that were included in the handout. The farmers toured the plot and worked on the questions in groups of three or four. One group answered 16 of the 18 questions correctly. All groups answered at least half of the questions correctly. Although a survey was not taken after completion of this workshop as had been done for the first workshop, many favorable comments were received both about the demonstration plot and how well it tied in with the first workshop and also about the visit to the corn breeding nursery. The latter activity was clearly a first for most of the farmers, and their interest was demonstrated by the large number of questions they asked.

3) Development of electronic lessons.
Four lessons were completed and are available at the University of Nebraska Croptechnology website (now the Plant and Soil Sciences eLibrary website; http://plantandsoil.unl.edu/croptechnology2005/pages/index.jsp). Furthermore, three of these lessons were submitted, reviewed, and accepted by the American Society of Agronomy journal, The Journal of Natural Resources and Life Sciences Education. The titles of the four lessons are
Corn Breeding: Lessons from the Past
Corn Breeding: Introduction to Concepts in Quantitative Genetics
Corn Breeding: Types of Cultivars
Corn Breeding: Mass Selection

Each of the lessons was written at the level of a beginner. That is, no prior formal training in genetics or statistics about the science of corn breeding is required. We did not formally survey the farmers to determine how often they used these electronic lessons. Based on our discussions with many of the farmers, it appears that most were not using the lessons. Our sense was that most preferred to learn from doing rather than reading. However, it may be that some of the farmers may use these lessons in the future for the purpose of review, particularly if some become directly involved in plant breeding activities.

4) Evaluation trials.
Seventeen of the 20 farmers expressed an interest in planting and harvesting an evaluation trial on their farm. Over 12,000 seed packets, each with two kernels, corn hand planters, and detailed planting instructions (Appendix B-1) were sent out to the farmers in early April, 2005. Also, seed was packaged for three additional trials that were planted on University of Nebraska land by the project leaders. In August, additional instructions (Appendix B-2), an electronic scale, data sheets for recording the data, and various harvesting supplies were sent to each farmer who had planted a plot.

Data were successfully collected from only 8 of the 20 evaluation trials. The reasons for the loss of the trials were varied. Five farmers chose not to plant the trials or abandoned the trials after planting, even though each had indicated an interest the prior fall in conducting a trial. Five other trials were lost to weather conditions, ranging from heavy snow after planting that caused poor germination to hail storms. Two other trials had to be discarded because of high levels of deer damage.

Losing some evaluation trials is quite typical, although the number that was lost in this project seemed higher than normal. Some of this could be attributed to inexperience of the farmers in planting this type of evaluation trial, but in our opinion much of it was attributable to bad luck. Was the loss of interest because of the amount of work that was involved with planting and harvesting a trial or the perceived amount of work? We attempted to contact each of the five farmers who either did not plant the trial after saying they would or who abandoned the trial after it was planted. We were successful in contacting three of these farmers. Two explained that they had to start working off the farm and no longer had the time. The third offered no reason other than he “was no longer interested.” From the recruitment stage, we were very open with the farmers about how much time we thought would be required to conduct the trials—approximately one-half day to plant and one day to harvest. No one indicated before the evaluations that this level of time commitment was beyond what they could do, although the ability to commit time unexpectedly changed for two of the farmers. Other than these two farmers, only one other indicated after the fact that the time commitment was too much. The time we estimated that would be required to plant and harvest a trial was quite accurate.

5) Final workshop and analysis of data from the evaluation trials.
Only two farmers committed to attending the final workshop that was planned for the spring of 2006. Therefore, the workshop was cancelled. Instead, the primary project leader analyzed the data and then in June sent a detailed report (Appendix C) of how the data were analyzed and interpreted to each of the 20 farmers. We contacted each of the five farmers whose data were used in this report approximately two weeks after it was sent out. Four had not yet looked at the report and one had looked at it “briefly”. We asked them to contact us after they had read the report to let us know their opinions—none did so.

We were disappointed by the decision of many of the farmers not to see the project through to the end and the apparent waning of their interest in this project. This was particularly surprising in light of the mostly very positive feedback obtained after the first two workshops and the indication by 17 farmers after the second workshop to participate in the evaluation trials. Because of class schedules, we were limited to scheduling the final workshop during the week of the University of Nebraska’s spring break. A few of the farmers clearly had conflicts. But it seemed that most simply had lost interest in the project. None of the farmers who had lost trials due to weather or animal damage expressed an interest in attending a final workshop.

Perhaps the occurrence of this waning interest should not have been too surprising. From the outset we understood that most farmers believe the seed companies are meeting their needs. The percentage that would have a strong interest in developing their own varieties is likely quite small, although nationally the actual number may be several hundred or higher. In retrospect, we should have advertised more widely and perhaps only invited those farmers who already had implemented an alternative cropping procedure. In our opinion, such farmers are likely to have a greater need and therefore interest in self-directed plant breeding programs. Also, they already have demonstrated a willingness to “step outside the box”. However, project resources did not allow for a much wider advertising campaign that we undertook. We further address this issue of low farmer participation in the last year of the project in the section on IMPACTS, ECONOMIC ANALYSIS AND FARMER ADOPTION near the end of this report.

The results and analysis of the data from the eight “good” trials in 2005 are explained in the detailed report sent to all the farmers (Appendix C-1) and are summarized as follows (all the tables mentioned below are in Appendix C-2):

i) Six of the trials were located in the eastern half of Nebraska and two in western Iowa. Both Iowa locations were non-irrigated, whereas four of the six Nebraska locations were irrigated. The final plant density at each location was close to 29,000 plants per acre, except for one non-irrigated Nebraska location at which the density was only 21,000 plants per acre. Data were obtained at each microplot trial on ear weight. Based on grain moistures that were estimated from the date of harvest and on the estimated weights of grain and cob, the mean grain yield at 15% moisture at the eight locations ranged from approximately 165 bushels per acre (at a non-irrigated, Iowa location) to 225 bushels per acre (at an irrigated, Nebraska location).

ii) Based on ear weights, the 312 topcrosses from NEL (hereafter these topcrosses are referred to as experimental hybrids) had relatively high performance compared to the 4 check hybrids. At each location, the average ear weight of the top 10% of the experimental hybrids was greater than the ear weight of the latest check hybrid (FR1064 x FR4310), which was the check with the greatest ear weight. This advantage ranged from 2% to 28% (Table 1). The values of this advantage may be biased upward because the experimental hybrids on average were a couple days later than this check hybrid and also were slightly taller.

iii) Based on within-location ANOVAs, highly significant differences were observed among the ear weights of the experimental hybrids in all locations (Table 2). The importance of this result is that it says the microplot protocol was effective in uncovering differences among the experimental hybrids. Based on estimates of broad-sense heritabilities (H) (Table 2), data of approximately equal value were obtained from each location.

iv) Based on across-location ANOVAs, the differences among the ear weights of the experimental hybrids, averaged across locations, was highly significant in 10 of the 12 tests and significant for the other 2 tests. In comparison, the experimental hybrid x location interaction was statistically significant in only 4 of the 12 tests (Table 3). Another way of considering the importance of the experimental hybrid x location interaction was to determine the frequency that specific experimental hybrids were among the best at multiple locations (Table 4). Twenty-one of the experimental hybrids ranked in the top 10% at three different locations, four ranked in the top 10% at four locations, and one ranked in the top 10% at five locations. The likelihood of these repetitive top performances based on chance alone is much smaller than what was observed. The important conclusion from these analyses is that for these experimental hybrids in this year the selection of the best hybrids for any one farm should be based on average performance across all locations.

v) Based on the probability of greater F-values from the within location analyses for the hybrid effect, there was no indication that the data from the microplots and the standard plots in 2005 differed in precision (Table 5). Based on H, there was some suggestion that two replications of data from standard plots were slightly superior to two replications of microplot plot data, but any advantage disappeared when the replications for the microplot protocol was increased from two to six.

vi) Correlations of hybrid performance for all pairs of trials, both microplot and standard plot trials, showed on average that two random microplot trials were not any more similar to each other than either was to a standard plot trial. However, the two earliest check hybrids, which also had the shortest plant heights, performed significantly better in the standard plots than in the microplots. This poorer performance in the microplots was least notable in the location with the lowest plant density. These observations are consistent with the poorer performance of these two check hybrids resulting from competition due to differences in plant height. Competitive advantages/disadvantages related to height would be expected to be greater in microplots than in standard plots and absent in pure stands. The conclusion is that if significant differences exist in height among hybrids being compared, then height should be taken when using microplots. Statistical analyses could be used to adjust the yield data for differences in height. Another option would be to replace single-plant plots with one-row plots of three to six plants. Only the inside plants in each row would be harvested. This would largely eliminate the competitive effects between adjacent plots within a row.

vii) The best 12 experimental hybrids across all 12 tests had an estimated mean grain yield that was 22.9% greater than and an estimated grain moisture that was 6.4% less than the mean of all the experimental hybrids (Table 6). Also, the mean yield of these 12 selections was greater than the yield of the best check hybrid and the grain moisture was drier. The estimate of broad-sense heritability (H) was 68.5% (Table 2). If the value of the narrow-sense heritability (h2) is one-half the this value of H (this is a reasonable guesstimate based estimates of genetic variances obtained in prior corn research (Hallauer and Miranda, 1981)), then the expected genetic gain in grain yield from this one cycle of selection would be 8%. This is comparable to previously reported genetic gains in corn selection studies done by professional breeders (Hallauer and Miranda, 1981).

6) Other activities—
i) The average grain yields at the three locations at which both microplot and standard plots were evaluated in 2006 ranged from 55 to 160 bushels per acre. The low yield of 55 bushels was observed at a dryland location and was less than one-half the lowest yield of any location in 2005.

Many of the trends observed in the 2005 trials were repeated in the 2006 trials. For example, with six replications of microplot data, highly significant differences were observed among the experimental hybrids at each location. With two replications of standard plot data, significant differences among the hybrids were observed only at the highest yielding location. Thus, six replications of microplot data were at least as discriminating as were two replications of standard plot data. Compared to 2005, there appeared to higher correspondence between the two types of plots with respect to which were the best hybrids. In one test, the same check hybrid was ranked first and an experimental hybrid was ranked second best with the microplot data and third best with the standard plot data. In a second test, three hybrids were ranked among the best four hybrids when using either the microplot or standard plot data.

ii) Seed of an improved version of the NEL population was produced by hand pollinations in 2006. However, because of extreme heat during the week of pollination, the amount of seed produced was too small to allow for this population to be released. Another seed increase is planned for 2007.

iii) A large balanced set of two-, three-, and four-way hybrids and also some various types of population hybrids were produced in the 2005 and 2006 nurseries. Evaluation of these hybrids will begin in 2007.

Research conclusions:

There is ample evidence that the farmers who participated in most of the activities of this project have an increased understanding of the principles of corn breeding. Also, the merit of a protocol by which the farmers can implement these principles to develop their own corn varieties has been established.

We did not do a formal economic analysis of the benefits a farmer would realize by undertaking his own variety development program. A cost benefit analysis would have no value until a reasonable estimate of the average annual genetic improvement that could be realized from such an effort by the farmer was available. Until this research was undertaken there was no evidence for even assuming than any progress could be achieved. Even now, although our work demonstrated that some progress appears likely, an economic analysis would be contingent on many factors beyond a farmer’s control and also would be highly speculative.

We suspect the speculative nature of outcomes from a relatively “small” self-directed plant breeding effort (small relative to the current activity of the private seed industry) was not lost on the farmers who participated in this project. This likely contributed to the loss of interest towards the end of the project and also may be a key reason why -— to the best of our knowledge -— none of the farmer participants has initiated a cultivar development program.

Since this project was initiated, several small companies have begun plant breeding programs to develop corn hybrids specifically adapted for the organic market. This is clear example of how the private sector will rapidly deploy resources to meet the seed needs of a niche market once that market attains a certain size and appears to be sustainable. It is quite natural, we believe, for farmers to have a keen interest in how the varieties they grow are developed. However, the incentive to become actively engaged in variety development may temporarily increase as the need for a new type of variety arises but then likely will fall as either that need is met by private companies or it diminishes.

This project was launched on a sincere belief that a strong possibility existed for the establishment of regional, farmer-directed plant breeding cooperatives or clubs. We now believe this is not likely in the foreseeable future, even though we have demonstrated probable improvements can be realized by such efforts.

Economic Analysis

(refer to “Impact of Results” section)

Farmer Adoption

(refer to “Impact of Results” section)

Participation Summary

Educational & Outreach Activities

Participation Summary

Education/outreach description:

Three significant and tangible outputs from this project have been completed or are in progress.
1) Four electronic lessons on corn breeding (listed on p. 7 under Results and Discussion). Three of these have been published in the Journal of Natural Resources and Life Sciences Education.
2) A research publication comparing standard plots and microplots is under development.
3) One farmer participant, as a result of his involvement in this project, became very interested in conducting replicated tests of commercially available organic hybrids on his farm. In 2006, we assisted him in designing this test and subsequently in evaluating the data obtained from this test. Also, we conducted a field day on his farm that attracted approximately 20 individuals. A write-up on how the test was designed and evaluated likely will be published in an organic newsletter with wide circulation in Nebraska.

Project Outcomes

Recommendations:

Areas needing additional study

Although we now believe that farmer-directed corn breeding programs will have very little impact in the foreseeable future, we still are concerned about the increasing concentration of corn breeding in a few large, private companies and the associated reduction in genetic diversity represented by the widely-grown corn varieties that likely will result from this concentration. Corn being the number one cash crop in the United States heightens this concern. We believe it is vital that public dollars be allocated to sustain long-term public corn breeding programs, particularly those focused on developing new germplasm sources. The Germplasm Enhancement of Maize (GEM) project is an example of public support in this area. Other public programs independent of this project should be supported.

A second area we believe requires support is the testing of privately-developed corn varieties. Many state universities conduct variety testing programs for corn and other crops. Most of these programs are financed by charging the participating companies entry fees. As the number of independent seed companies declines, the economic sustainability of these programs is becoming challenged. Yet, the farmer need for independent data that can be used to select varieties remains strong. Also there is a strong need for farmer education in how to evaluate these data. We suspect future independent testing will involve partnerships among universities, various farm groups, and large buyers of corn seed, such as farm management firms. These partnerships will need to become largely economically self-sustaining, but some public dollars will be essential to support their formation.

Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.