Heritage Maize Project

[Editors Note: The report for this project includes tables and appendices that could not be included here. For a copy of the entire report please contact North Central Region SARE at ncrsare@umn.edu or 1-800-529-1342.]

ABSTRACT
The historical claim of higher food value in Open Pollinated (OP) corn was examined, along with claims of potential economic equality between modern hybrid corn varieties and OP corn varieties. Our initial survey of OP corn from across the cornbelt and parts east found that many OP corn varieties were higher in protein, some were higher in fat, and a few were higher in other mineral constituents and lysine than commercial hybrid corn grown nearby. Our initial strip trial with several OP varieties in southern Wisconsin demonstrated poor stress tolerance and yield by OP corn under a drought in a corn-on-corn field. Additional data from field evaluations carried out by Iowa State University, Cornell University and the Michael Fields Agricultural Institute demonstrate that the OP varieties tested are generally not competitive for yield with modern hybrid varieties though they are often a percentage point or more higher in protein. Using CornPro3, an economic prediction tool that we put together with Mike Duffy and Anne Holste of Iowa State University and Tom Tylutki of Cornell University, we found that the increased feed value and lower seed costs of OP corn varieties do not at this time make them competitive with F1 hybrids. However, preliminary data suggest that one of the university breeding populations could make a much improved OP variety that could be economically competitive with many current hybrids. For an OP variety to be economically competitive and not unduly increase the number of acres needed to feed animals, a yield of 90% of current hybrids would be necessary along with slightly higher protein levels.

INTRODUCTION
In 2000, we three farmers decided to find out the degree to which open-pollinated (OP) corn was more nutritious than commercial hybrids. We also set out to determine if organically grown corn was more nutritious than corn grown with chemical-intensive methods. Our concern with nutrition was based on the idea of growing a cheaper feed that might be more economically valuable for farmers who feed their own animals. Stories about the feed value of OP corn abound and we wanted to see if this was so and if it was, would it matter.

The North Central Region Sustainable Agriculture Research and Education Program elected to fund our study. We persisted with it even with drought in 2001 and one of us taking up corn breeding at graduate school. We had many grain samples analyzed and have also encouraged more evaluation of OP corn at several universities. We have amassed data on OP corn from these university programs and from farmers. With university cooperators we also developed a computer tool for carrying out the economic analysis of grain costs when feeding two species of livestock. Our report is later than we had wished, but the additional time has allowed us to analyze a more meaningful set of data in the determination of the value of OP corn.

METHODS
In 2000 we asked farmers to send to us samples of their OP corn and also commercial hybrid corn from nearby fields. We obtained a good sample of OP varieties from across the country, though a much smaller sample of commercial hybrid grain. These samples were analyzed by the Stearns County DHIA lab in Minnesota using wet chemistry methods. They were analyzed for energy, fat, protein, major minerals, and lysine. Troy Salzer, our cooperator with the Minnesota Extension office in Carlton County, found that the higher protein in OP corn would make its feed more valuable by saving farmers some of the expense of protein concentrates (See appendix).

In 2001 we asked for a similar set of samples from OP corn growers but received limited cooperation this time around. A strip trial set to be planted in Minnesota was apparently not planted or at least we quit hearing from the cooperating farmer. A strip trial in Wisconsin was planted and we held a field day there in the fall. Margaret Smith of the Cornell University corn project evaluated OP, synthetic and hybrid varieties at several locations in New York. Kendall Lamkey from Iowa State University and Walter Goldstein of the Michael Fields Agricultural Institute in southern Wisconsin carried out a wide ranging evaluation of corn variety types from which data were provided to us. Our findings were written up for a publication put out by organic farmers in Kansas.

In 2002 interest in continued strip trials vanished since the results in 2001 were so poor. We continued to get data sent to us from cooperating universities and from the Michael Fields Agricultural Institute in Wisconsin. The Cornell University corn project provided us with samples from ongoing evaluations of OP varieties and a student project on organic corn production (sponsored by the Organic Farming Research Foundation) also provided us with yield and nutritional quality data. From here on out we employed the much cheaper and still fairly accurate near-infrared spectrometry methods for grain quality analysis. Some of our results were presented at the ASA, CSSA, SSSA meeting in Indianapolis that fall.

In 2003 we continued to take in samples for analysis from the Cornell University corn project and to develop a computer modeling tool for an economic analysis of the value of higher protein in corn and the utility of OP corn. Our cooperators at Iowa State University, Mike Duffy and Ann Holste, put most of the finishing touches on the spreadsheet tool while Tom Tylutki of Cornell University provided many useful suggestions on how to find information on the formulation of feeds for livestock. This spreadsheet tool, CornPro3, was finished in the summer of 2004 and our final analyses carried out in early fall.

CornPro3 has some underlying assumptions that must be addressed. First, it is not being used to evaluate the total cost of feeding animals but is used to determine the costs of feeding grain (corn and soy meal in this case). The tool has places for the entry of yield and crude protein values for any variety, but is set up specifically for evaluating the average or the best performing OP variety, synthetic variety, and check hybrid. It has built in assumptions about the costs and management of these types of varieties and all of them can be readily changed. There are places for entering the seed costs and seeding rates, and on a separate page there is a place to enter the costs of production based initially on Mike Duffy's 2004 economic assessment for producing corn in Iowa. The outputs are graphs of the costs for feeding grains to 100 head of livestock for one day and for a year.

The rations for a feeder steer, a lactating holstein cow, and for small and finishing pigs are based on rations found in the most recent editions of National Academy of Sciences publications concerning the nutrition of these animals. These diets are used for the sake of simplicity since they are easily found in these books. However, they may not represent most rations and any farmer should carefully consider his or her own rations and make modifications if they vary much from these "standard" rations. The only concentrate used for additional protein in this model is soy meal. The cost for soy meal has varied widely in the last year, so most of our results are with high soy prices ($300/ton) and more recent prices ($200/ton). High soy prices favor the economic value of higher protein corn varieties since these usually provide cheaper protein.

The seed costs on the combosheet (where any of the four rations and several other analyses can be observed) are based on averages for purchased seed. Some of our results are with the average hybrid seed cost (taken to be about $100/bag) though we have also made comparisons with cheaper ($60) and more expensive ($140) hybrid seeds. OP seed corn is priced at $40 per bag since this is a common price and about average for what we have seen. The costs for synthetic variety seed are higher in case the universities would want a royalty on this seed, though for the time being we are not aware of any university synthetic populations actually being commercially available. If one were to estimate the seed costs for saved seed this would be substantially lower than for purchased seed. In our synthetic variety modeling table we estimate seed costs for saved seed at $20 per bag to cover the opportunity costs of time spent selecting good ears (1-2 hrs), drying, shelling and cleaning. This is an estimate that favors the hybrids since usually we farmers fully discount this time and consider this seed as "free." By using $20 per bag we feel we are being very fair to the actual cost of obtaining good saved seed.

Our general model is based on lower planting densities for OP and synthetic varieties because this is usually how they are planted. Most OP corn growers recommend 18,000 plants/A and most hybrid growers are planting for 28-32,000 plants/A. This estimate favors the OP varieties by making them cheaper to plant per acre, however in our synthetic variety modeling tool we use the same planting density as the hybrids.

In the costs of production tables we use Mike Duffy's cost estimates directly. These cost estimates suggest that the expenses in producing corn are very high these days. High costs of production make high yielders more economical since high yield lowers the cost of producing each bushel. However, we do use estimates for some machine time, drying, storing and shipping costs that vary with the corn yield. This favors the OP varieties slightly since they tend to yield less, but it is true that less corn would have to be dried or shipped. We assume that less fertilizer would be needed for lower yielding varieties based on the amount of fertilizer carried off with the harvest and have adjusted fertilizer rates accordingly. We are not sure if one could get crop insurance for a field of OP corn. However, we have kept that cost of production in for the sake of simplicity. This may slightly favor the hybrids. In the synthetic modelling we held fertilizer equal across all yield levels for this analysis.

While it is true that all OP varieties are not equal this is true of any corn varieties. One usually goes with the better ones. To simplify our analyses, we calculated an average OP corn yield and protein content for the two best looking varieties, Nokomis Gold released by Walter Goldstein of the Michael Fields Agricultural Institute in Wisconsin and “Wapsie Valley” released by Victor Kucyk of Dublin, ON. We now have access to many years and locations worth of data on these two varieties from Iowa State University, the Michael Fields Agricultural Institute and Cornell University. We did the same for the hybrid checks that were used wherever these varieties were evaluated, removing those that clearly were not adapted. These averages are weighted by the number of locations in each year the varieties were evaluated. Since these OP varieties and their hybrid checks so often performed similarly across the data sets, we averaged them and present in the model the average yield of these two OP varieties and of the hybrid checks used in these various studies. A preliminary analysis for synthetic varieties based on a much lower number of evaluations was also carried out and that average entered into the model. Especially low yielding hybrids, OP varieties, and synthetics were not used. The disaster we had in 2001 was also not used in this average since we have plenty of years worth of droughty conditions from New York and since OP growers almost never grow corn on corn. We also have foregone delaying the project further waiting for the incoming 2004 results in New York which almost certainly reflect a bumper crop, even for the OP varieties. A strip of left over OP corn seed in Wisconsin also looked relatively good in these wet conditions. We do not feel that the inclusion of these data would in any way change our basic observations or conclusions.

RESULTS
In Wisconsin in 2001 the hybrid check in our strip trial yielded about 100 bushels/A during the drought at that location with 3% lodging. The three OP varieties (Reid, Nokomis Gold, E-95) yielded 15-20 bushels/A with an average of 50% lodging. Small plot observations of several OP varieties in Kansas provided similar low yields and many barren plants. In New York, late maturing Greenfield 114 yielded 60% of the hybrid average at one location and 30% at another. Lancaster and Silver King yielded much less and all three had much higher rates of lodging than the hybrids. Most midseason OP varieties yielded below 100 bushels/A while the hybrids yielded 151-194 bushels/A and stood much better. The 1-3 points of higher protein generally did not appear to be of much value with such low yields for most OP varieties.

Over the years that we have been amassing data from trials carried out in the Midwest and in New York and New Hampshire, the OP varieties have occasionally been competitive with some of the check hybrids, but usually this has not been the case. OP yields have usually been in the 54-70% of hybrid checks range with some far less (see appendices). They have often had higher protein than did the hybrids but this has not always been the case (see Table 1). Most of these data are presented in the Appendix at the end of the report and come to us from university collaborators and from the Michael Fields Agricultural Institute.

The most recent data and corn to come our way are from 2003 trials carried out by the Cornell University corn project. We had the samples analyzed and were provided the agronomic data that come to us from two experiments in three locations. The first is a comparison of Cornell hybrids with hybrid checks, synthetic varieties and an OP variety (“Wapsie Valley”) in one organically managed field. The OP variety yielded as well as the synthetics but yielded less than the experimental hybrids (Table 1). The amount of protein in the varieties did not vary much, but the second highest % crude protein was actually from one of the experimental hybrids.

The second experiment (Table 2) was conducted at three locations and included several of the OP varieties that our initial study indicated were high in protein (CG-Cuzco, Iroquois White, Ranger). These varieties did not prove to be as high in protein as our initial analyses suggested, though all but the CG-Cuzco were higher than most of the hybrid checks. However, yields were generally very low. One very interesting result that surely deserves further attention comes from one of the fields that was organically managed and where they could not plant the treated hybrid checks. In that organic field they used an NC+ brand hybrid variety and a synthetic from Liz Lee of the University of Guelph (CG-Stiff Stalk Cycle 5). The hybrid variety yielded 173 bu/A while the synthetic yielded 164 bu/A.

Our observations and experiences, and also those of university researchers, have found two OP varieties to be generally superior to the rest for grain yield and standability. Average yield for these two best performing OP varieties (Nokomis Gold and “Wapsie Valley”), weighted by locations within years, are presented along with the means of their hybrid checks (Table 3). We had protein results in only half of the locations, but these are the best results we could obtain and they cover many years and locations for a more realistic estimate of yield and protein content. “Wapsie Valley” and Nokomis Gold yielded about the same, 82 bu/A, and the hybrid varieties also generally yielded just over 130 bu/A. All further economic analyses for current OP varieties are based on the averages of these OP varieties, hybrid checks and the preliminary data for synthetic varieties from Table 3.

Our basic economic evaluations are presented in Figures 1-6. In all cases the cheapest grain for feed rations was grown with an F1 hybrid whether that hybrid cost $60 per bag (Figs. 1 and 2), $100 per bag (Figs. 3 and 4) or $140 per bag (Figs. 5 and 6). This remained true whether soy prices were $200/ton (Figs. 1, 3, and 5) or $300/ton (Figs. 2, 4, and 6). Clearly the most important factor in feed costs was the yield of the variety since this value is used to determine how much of the costs of production accrue to each bushel of corn. This cost/bu difference is obvious in all the ration costs across species: with high costs of production low yielding corn is expensive to feed. Secondarily, seed costs and protein content had smaller economic effects. For instance, one point of extra protein in the OP varieties allowed the feed costs for 100 beef steers for a year to go up about $4200 when soy prices rise to $300/ton, whereas the hybrid corn feed costs go up about $4800 (Figs. 3 and 4). This $600 difference is much smaller than the roughly $10,000 difference in feed costs that already existed between using the hybrid corn and the OP corn. Similarly, cutting back the seed expenses of the OP corn to reflect only the opportunity costs of producing one's own seed ($20/bag) only drops the costs for feeding grain to those 100 steers by about $600 in a year. With these current conditions, yield and costs of production are the primary economic considerations for feeding corn, followed by seed price and protein content of the grain. The cost of seed independence is currently fairly high.

Could a variety from which farmers saved seed, whether OP or Synthetic, ever be economically competitive with F1 hybrids? To answer that question we used the third modeling tool from the combosheet of CornPro3. That model looks at the cost of feed for varieties that are planted at the same rate as hybrids and raised in the same way as hybrids, but for which seed costs are low and yield and protein content vary. With soy prices held at $200 per ton, one can see that at $60 per bag of hybrid seed the F1 hybrid is the most economical for feeding on farm (Table 4). Compared to $100 per bag hybrids the OP/Synthetics are economically competitive when they yield 90% or more of the F1 hybrid average and for $140 per bag hybrids the OP/Synthetics that yield 90% of the hybrid average are economically superior. These differences are entirely due to the difference in seed cost.

When protein is considered, the 90% OP/Synthetics are economically comparable when they are three points higher in protein than the $60 per bag F1 hybrids. When soy is high equivalence is reached when the 90% OP/Synthetics are two points higher in protein. The OP/Synthetics that achieve 80% of the F1 hybrid yield are equivalent in value only if soy prices are $300 per ton and they have four points of crude protein over the F1 hybrids.

When hybrid seed prices are $100 per bag, one point of higher protein makes the 90% OP/Synthetics economically comparable. For the 80% OP/Synthetics they are comparable in feed cost when they have four points higher protein than hybrids when soymeal is $300 per ton. When hybrid seed prices are $140 per bag, the OP/Synthetics that yield 90% of the hybrid average are always economically superior regardless of the amount of protein. The 80% OP/Synthetics would be economically comparable to F1 hybrid varieties for feed prices when they were five points higher in protein at $200/ton soy prices and when there were three points higher in protein at $300/ton soy prices. With our model and these inputs, varieties yielding less than 80% of hybrids cannot achieve economic equivalence for feed prices unless they had extremely high protein levels that were still useful to animals. This is highly unlikely given research results from the 1940s and 1950s.

What would happen if we lowered these costs of production, say by using low input/organic methods? In all cases the costs of grain for animals fed on farm drop dramatically when low input methods are used (Table 4). If rents are dropped along with a shift to low input methods the feeding costs drop even further. As these costs of production are dropped the economic viability of feed from lower yielding varieties increases. It appears that as costs of production drop the impact of seed price increases. OP/Synthetic varieties yielding 80-90% of the F1 hybrid would provide comparably priced or even cheaper feed when grown under low input methods than would the F1 hybrids raised under conventional methods. HOWEVER, F1 hybrids are competitive or superior in lowering feed costs in these low input scenarios if the OP/Synthetic yield drops below 90% of the hybrid yield, unless the lower yielding varieties have much higher protein than the hybrid.

Another interesting way to use the OP/Synthetic modeling tool is to use it to predict the economics of using hybrids that may yield less under low input/organic methods than they would under conventional production methods. (Our data initially showed a protein advantage to using organically grown corn, though subsequent analysis of samples from replicated university trials did not bear this out.) Using our estimated low input COP and switching all seed costs to $100, corn could lose up to 10% of its yield under conventional methods and still be economically superior for producing feed. Costs for feeding 100 head of steers, dairy cows or finishing pigs drop to $33023, $42379 and $11298 when COP drops and the same F1 hybrid (132 bu/A) is used. If COP were reduced further even a drop of 20% of yield could be economically competitive with the same variety produced under conventional management. What if the corn was for sale instead of fed to animals on farm? With the second spreadsheet tool on the combo page we found that when corn is $2.50 per bushel there is no profit per acre unless low input methods are used and hybrid seed costs are low (Table 5). By examining the effects of lower yield we found that no lower yield situation was profitable when corn was for sale. Losing 10% of our 132 bu/A yield in a low input scenario lost less per acre than did 100% yields at high input costs of production. Still, net loss is net loss and profitability is required for sustainability. Keeping costs down and yields up appeared the only way to profit by selling corn.

Finally, while we are unable to predict whole farm profitability given the yields and feed costs that we have predicted, we were able to model the amount of land needed to feed 100 head of animals for a year given the different rations, yields, and protein concentrations. The second yellow column in the OP/Synthetic modeling tool of CornPro3 shows how many acres it would take to produce the grain for 100 animals receiving the rations for a year. While economic equivalence or superiority was found for some low yield situations, more land was usually needed to feed the animals when corn yields were lower (Table 6). To have land use equivalent to current hybrids, OP/Synthetics would need to have at least 90% of the normal hybrid yield and higher protein to reduce the acreage of soy beans. With lower yielding varieties the number of animals that could be fed on a given farm would be lowered, thereby capping the potential gross income from the farm. Another obvious observation from these data is that higher protein hybrid corn could reduce the acres needed to feed animals from current levels if the protein in the corn were taken into consideration.

DISCUSSION AND CONCLUSIONS
Repeatedly we found that OP corn tends to have higher quality as farmers have been saying for decades. This is not always the case, but it is a strong and often observed phenomenon. Results across locations have shown that the differences are often small. Of the differences we observed, protein content appears to be the most repeatable and important difference. In some cases there was a higher fat content in the OP corn. Otherwise, the grain quality of F1 hybrid and OP varieties appears fairly similar: there is more variation among varieties than between these types of varieties.

The two OP varieties that regularly top the university and NGO evaluations in recent years, Nokomis Gold and “Wapsie Valley”, yielded substantially less grain than did hybrid checks in evaluations across the northern corn belt. Their protein content was only one point higher than the average of the hybrid checks and this made only a little economic difference in their favor. These two varieties of OP corn have been found to be competitive with hybrids for silage production in the SARE trials in New York and New Hampshire. However, under these Iowa economic conditions, we have found them not to be economically competitive for grain production even when that grain is fed on farm.

Compared to yield, much lower seed cost and the savings from higher protein are small improvements in the economics of growing corn for feed when OP yields are this low and the differences in protein this small. At this time and with our data we cannot recommend growing OP corn for grain based on its economics, which are mostly driven by low yield and higher production costs/bushel. We have defined under what conditions an OP or Synthetic variety would be competitive or superior to F1 hybrids for producing lower cost feed under Iowa costs of production. Low input management would also help a great deal. However, the current OP varieties for which we have data do not come close to yielding the 80-90% of the hybrid checks needed to be competitive under the economic conditions modeled in CornPro3. The levels of protein needed to help make low yielders more competitive are also rarely observed, though a 2-3 percentage point increase is possible. The preliminary data we have on synthetic varieties from the universities also generally show yields that are 70% of hybrid yields or less.

Clearly, developing an economically competitive OP corn for grain production would be a breeding challenge. The only variety we have seen that appears to have any chance of coming close to economic competitiveness with F1 hybrids is the synthetic from Guelph, CG Stiff Stalk C(5), though the data we have for it are from only one field and one year. Liz Lee at Guelph reports average yields for that variety of 118 bu/A and our hybrid average yield was about 132 bu/A, therefore we recommend further strip trial comparisons of that variety with F1 hybrids in the northern corn belt. Should this variety prove to be economically competitive for producing feed we recognize that with its protein levels about 7% more acres would be needed to feed animals. The overall economics have yet to be determined.

Our model and analyses are open to criticism. We have attempted to be as realistic and fair as possible, but have had to make some assumptions that require validity testing. We do not know whether all users of low input methods could maintain this same average yield. We do not know whether OP varieties could produce what they do with lower nitrogen applications. We do not know the degree to which cheaper feed would actually be profitable if more acres are needed to feed animals, though a yield of 90% of the hybrid yield seems to have potential. One of the assumptions made is the degree to which higher protein in corn would actually substitute for soy. Research from the 1940s suggested that crude protein above 13% was not useful as most of that additional protein was of the low digestibility, low quality zein fraction. This still gives a wide range for consideration (7-13%) and increases of several percentage points are being reported from the Genetic Enhancement of Maize Project. We hope that current efforts to increase protein in corn will be carried out with animal scientists who can test the economic and biological utility of raising protein in corn.

Perhaps the most useful finding in the short term is the utility of lowering costs of production and the potential of low input methods for improving farm profitability. When corn was sold only low input methods and low cost seeds produced a profit; conventional management (based upon many averages) always produced net losses at this yield level. Cheaper seed can be very helpful if yields aren’t reduced too much. We hope that farmers and others will use our model and adapt it to their needs to help them address the economics of corn production and to recheck our findings. We are making CornPro3 available to assist them in doing so. Just email us at fk29@cornell.edu to get a free copy of this Excel spreadsheet via email.