Evidence of the Magnitude and Consequences of the Roundup Ready Soybean Yield Drag from University-Based Varietal Trials in 1998
Over half the soybeans planted in the United States in 1999 are varieties genetically engineered to tolerate applications of the broad-spectrum, systemic post-emergence herbicide glyphosate manufactured by Monsanto Company (Trade Name, Roundup). Just a small fraction of soybeans produced in 1996 were "Roundup Ready" -- varieties able to tolerate direct applications of glyphosate.
The rapid adoption of Roundup Ready (RR) soybeans has been unprecedented in the history of American row-crop agriculture. No new genetic trait, nor any pesticide has so dramatically gained market share in such a short period of time. Roundup Ready soybeans have proven so popular with farmers because they greatly simplify the task of managing weeds and help farmers avoid a variety of problems associated with other herbicide-based weed management systems. They have proven especially popular among farmers who must complete weed management practices on a timely basis over hundreds to thousands of acres of soybeans.
The success of RR soybeans is remarkable in light of the magnitude of the so-called Roundup Ready "yield drag." Under most conditions extensive evidence shows that RR soybeans produce lower yields than possible if farmers planted comparable but non-engineered varieties.
This report reviews the results of over 8,200 university-based soybean varietal trials in 1998 and reaches the following conclusions regarding the magnitude of the RR soybean yield drag -
It is important to place the RR soybean yield drag in perspective. From 1975 to 1994 soybean yields rose on average about 0.5 bushels per year. In 1999 the RR soybean yield drag could result in perhaps a 2.0 to 2.5 percent reduction in national average soybean yields, compared to what they would likely have been if seed companies had not dramatically shifted breeding priorities to focus on herbicide tolerance. If not reversed by future breeding enhancements, this downward shift in soybean yield potential could emerge as the most significant decline in a major crop ever associated with a single genetic modification..2
On whether RR soybean systems reduce pesticide use and increase grower profits, our analysis shows that -
The remarkable popularity of Roundup Ready soybeans, despite their cost and the significant yield drag associated with their use, is evidence of the difficulty and high cost of today's herbicide-dependent soybean weed management systems. The rapid evolution of weeds better able to withstand applications of Roundup reinforces the need for more integrated, multiple tactic weed management systems.
A. Soybean Weed Management - A Never-Ending Challenge
Year in, year out dealing with weeds is the toughest pest management problem facing soybean farmers. Perhaps the only tougher challenge facing growers is finding a way to make a decent profit, especially in years like 1999 when surplus supplies and soft demand depresses prices.
Dr. Donald Wyse, a weed scientist at the University of Minnesota, highlights the dominant role weeds play in shaping Midwest row-crop farming systems --
"Weeds are the major deterrent to the development of more sustainable agriculture systems...[they] dictate most of the crop production practices (e.g., tillage, herbicides, cultivation, row spacing)...It must be emphasized that current cropping systems enhance weed populations, forcing farmers to use tillage or herbicide inputs to manage weed problems." (Wyse, 1994)
Each year farmers have to deal with a few to a half-dozen species of grass weeds, and another few to half-dozen broadleaf weeds. It's not a question of whether weeds will pose a threat to yields, but which ones and -
A poor weed management job in one year can make matters worse for years to come. It takes years of careful attention to detail in weed management for farmers to reduce the number of weed seeds in the soil profile enough to make a significant difference in the weed pressure in subsequent years. Even then, some weed seeds can move great distances with the wind, so no farmer is in complete control of his or her weed management future.
Poor weed control costs farmers money and pride. There is nothing more visible in a soybean field than a tall patch of foxtail or velvetleaf wafting in the breeze above the crop canopy. When it comes time to harvest a field, patches of weeds can slow down the combine and lead to uneven maturation and drying of the crop. Weeds also compete with the growing crop for sunlight, water and nutrients. The greater the number of weeds, especially early in the season, the bigger the threat to yields.
Little Real Progress
Weeds pose roughly as much trouble for farmers in 1999 as they did in 1959, despite billions in expenditures on herbicides and hundreds of millions more invested in research, development and registration of new herbicides (Benbrook et al., 1996). Why? Because farms have gotten bigger and more specialized and many changes in agronomic practices have actually tipped the playing field to the advantage of certain weeds. To reduce soil erosion losses and to cover large acreage fast with minimal demands on human labor, conventional farmers have all but abandoned some proven, non-chemical weed management tactics. For 40 years the USDA and soil conservation advocates have supported the trend toward conservation or reduced tillage.
Primary tillage is carried out to kill or damage perennial weeds, bury weed seeds and prepare a seedbed but it also exposes the soil surface to the erosive effects of rainfall and wind, a disadvantage on highly erodible land. Tillage with lighter equipment, called cultivation, is carried out after crops have emerged to help manage weeds. Cultivation is an effective, low-cost way to suppress weed populations and remains the back-bone of non-chemical weed management systems. Yet for a variety of reasons farmers have moved away from even this sort of tillage. One of the major reasons is the trend toward narrow row planting systems, solid seeded fields and no-tillage systems.
Another is that pulling steel across fields is expensive. It takes a lot of fuel and tillage equipment requires constant repair and maintenance. Plus, large machines are needed to cover large farms quickly with limited labor inputs. The bigger the machine, the more capital tied up and the greater the pressure to cover more ground. It's a vicious circle that tends to lead farmers progressively away from multi tactic, non-chemical weed management systems.
Integrated Weed Management systems are proven and profitable alternatives to systems largely dependent on herbicides (Tillman, 1998; Drinkwater et al., 1998; see also multiple research reports from the Aldo Leopold Center for Sustainable Agriculture at Iowa State University and annual progress reports from the Practical Farmers of Iowa, see http://www.pfi.iastate.edu/ ). They can reduce reliance on herbicides and lower costs with no loss of yields but they take planning, skill and commitment. Key components for success include a proper rotation, tillage and planting systems designed to get crops off to a fast start, and mechanical cultivation to combat weeds when they are small and most vulnerable. Success depends on developing a good understanding of the biological processes and cycles unfolding on the farm. Careful timing and considerable skill in the integration of multiple tactics is essential.
The human dimension is actually perhaps the dominant constraint. Individuals with the skill and experience needed to properly operate cultivation equipment do not yet qualify for the endangered species list, but their numbers are falling fast with no end in sight. Money to pay wages is not the only issue since weed management costs have clearly rising over the last 15 years. The $15.00 increase in per acre herbicide.5 expenditures on most farms would pay for two to three additional passes with a cultivator.
Herbicides are the primary weed management tactic on 95 percent plus of the land growing soybeans because it has become easier to get the job done with herbicides than with multitactic systems. It has become easier because of years of public and private investment in knowledge and infrastructure dedicated to the delivery and application of herbicides, and years of relative neglect to research, education, and technology development dedicated to improving the ease and efficiency of non-chemical weed control practices and systems. Priorities have not shifted in a vaccuum. Heavy advertisement has reinforced the notion that "progress" equals applying the latest herbicide.
But even without much support from the private sector or public research programs, farmers practicing sustainable methods or producing for the organic market have continued to refine non-chemical and reduced chemical weed management systems and have cut the costs of weed management to half or less of some neighboring farms using herbicide-based systems (see the sections on weed management in Pest Management at the Crossroads, Benbrook, C. et al., 1996).
Weeds have great capacity to adapt and evolve. Two Iowa State University weed scientists, Dr. Bob Hartzler and Dr. Doug Buhler, have studied one reason -- weed emergence patterns. In 1998 research, they found that different weeds emerge at different points during the season.1 In Northeast Iowa, fall panicum first emerged 50 days after initial emergence of giant ragweed. Across the state, there was almost a two-month span during which new weed species emerged.
The long period during which different species emerge for the first time is a special problem for farmers planting Roundup Ready soybeans. This is because Roundup works only on growing weeds. Farmers must wait for weeds to emerge to apply Roundup. But if they wait too long, weeds can get too big of a jump and some will only be stunted by Roundup application, requiring a second treatment or application of another active ingredient. Making matters worse, several weed species can germinate for a month or more, including some of the most difficult to control late season weeds like waterhemp and morningglory. As a result, a properly timed early season application may miss soon-to-germinate weeds that can then cause problems later in the season.
1 For more information, see the extensive weed management resources on the Iowa State University Extension weed management site at <http://www.weeds.iastate.edu/mgmt/qtr98-4/emergencepatterns.htm >.
Evolutionary changes in weeds have steadily pushed farmers to seek out "stronger medicine" in dealing with annual problem weeds. As a result, expenditures on herbicides doubled in the 1980s, rising from around $13.00 per acre in 1988 to over $26.00 per acre (see "Table 1, Soybeans - Seed and Chemical Production Expenses in Corn Belt States, 1975-1997," Benbrook, 1999). Competition-driven price reductions starting in 1998 are likely to keep expenditures on herbicides from rising for the next few years despite increased reliance and use.
Weed shifts and the emergence of tolerant and/or resistant weeds are the dominant mechanisms likely to undermine the long-term effectiveness of any herbicide-based system that is relied upon too widely or intensively.2 Weed shifts reflect a change in the relative proportions of different weed species present in a field. The emergence of weeds resistant to a given herbicide, on the other hand, is a genetically-driven evolutionary phenomenon which occurs within a species present in a field.
In the case of Roundup Ready soybeans, evidence has emerged already that such shifts and natural adaptation are occurring. Iowa State University researchers have documented glyphosate resistance in waterhemp species (see footnote for source). This particular species of weed is highlighted in the 1998 edition of the Weed Control Manual as the most worrisome "Up and Coming Weed" both nationwide and in the North Central region (Meister, 1998). Reasons why include the emergence of resistance to triazine and ALS herbicides and its ability to withstand Roundup applications (for an up to date overview by Iowa weed scientist Bob Hartzler, see <http://www.weeds.iastate.edu/mgmt/qtr98-4/roundupfuture.htm>).
Soybean farmers in the Midwest are also reporting that velvetleaf and smartweed species are tolerating application rates of Roundup that were effective just last year. Over the next few years, weed scientists anticipate that the gradual shift toward weed species that can tolerate Roundup will continue, requiring farmers to change their mix of practices. If even a small percentage of growers respond instead by just increasing rates of Roundup application and/or spraying more often, this could generate enough selection pressure to lead to the emergence of genetically resistant strains of commonplace weeds. In all likelihood these strains would quickly spread and all growers would then have to switch to other herbicides and control tactics.
Ironically the remarkable success of Roundup Ready soybeans in the marketplace sets the stage for problems in the field. Regardless of how well a given herbicide or weed management system is working, tomorrow's problem-weed is just around the corner. The more uniform the weed management system in an area, the quicker new problem-weeds are likely to emerge. The greater a farmer's reliance on herbicides, especially one or a few products, the more likely it becomes that resistance weed species will emerge. These are the lessons of weed management history that are not being adequately heeded as the biotechnology revolution reshapes the American soybean industry.
2 For a good overview of these potential risks, see Dr. Mike Owen's 1997 review of herbicide-tolerant technology (Owen, 1998).
B. Methods to Compare Yields
Two major methods are used to compare the yields of Roundup Ready soybean varieties to the yields of non-engineered soybeans. One involves comparing average yields across many different trials, or across many farms. The problem with such comparisons is that many variables are not controlled. Average yield differences might be explained by genetic difference or by other factors.
The most accurate method is to compare yields in properly conducted side-by-side trials carried out with near-isogenic lines 3 that differ only in the possession of the glyphosate-tolerance gene. Such trials are the most reliable way to isolate the consequences of genetic differences, all other things being equal.
3 By "near identical isogenic lines" we mean two varieties that are nearly identical in terms of their genetic makeup, except for the insertion of a specific gene or trait.
Most land grant universities in states with significant acreage of soybeans carry out soybean yield trials to provide farmers with impartial, accurate data on the performance of different varieties. (Internet-accessible sources of land grant university soybean trial data are listed in the references). The trials are typically conducted in different regions of a state and cover different maturity groups 4 . In some states, trials are carried out using a variety of tillage and planting systems.
4 Soybean varieties are often sold and studied within "maturity groups." In general, the longer the maturity period for a variety, the higher the expected yields. It is therefor inappropriate to compare the yields harvested of a long season variety in contrast to a short season one.
In this report we compare soybean yield trial data from university trials, the most impartial data available. We have drawn heavily on a University of Wisconsin analysis of over 8,200 soybean yield trials carried out in eight states in 1998. This definitive assessment of the Roundup Ready soybean yield drag was done by Dr. E. S. Oplinger, Dr. M. J. Martinka, and Dr. K. A. Schmitz and is entitled "Performance of Transgenic Soybeans in the Northern U.S." 5
5 This paper is accessible online at http://www.biotech-info.net/soybean_performance.pdf . The summary table from this paper is accessible at http://www.biotech-info.net/yield_performance.pdf . A slide presentation by Dr. Oplinger further summarizing their findings is accessible at http://www.uwex.edu/ces/soybean/slides/1998%20ASTA%20Expo/sld001.htm .
State Varietal Trial Results
In Tables 1 and 2 we report the differences in the yields of Roundup Ready ("RR") in contrast to conventional ("Conv") varieties in eight states, drawing on the data reported in the Oplinger summary report. Each table offers three comparisons - trial means across all varieties, the mean of the top five varieties tested, and the top varieties tested.
Table 1 reports bushels harvested per acre and Table 2 presents the RR yield drag in two ways - the difference in bushels per acre between the Roundup Ready varieties and conventional varieties, and second, as a percent of the yield of conventional varieties.
Oplinger data show that the -
Table 1. Roundup Ready (RR) Soybean Yield Compared to Conv. (Conv) Varieties in Eight States, 1998 [Bushels per Acre]
Top Five Trial Mean Varieties Top Variety States Conv RR Conv RR Conv RR Illinois 58 60 65 65 67 67 Iowa 61 57 64 60 66 60 Michigan 66 64 74 69 78 70 Minnesota 66 61 73 67 74 69 Nebraska 58 51 65 58 66 60 Ohio 60 58 67 63 69 65 South Dakota 49 44 54 50 56 51 Wisconsin 71 69 85 82 87 84 Source: 'Performance of Transgenic Soybeans - Northern U.S.', E.S. Oplinger, M.J. Martinka, K.A. Schmitz. For details by State, see Table 1. Yield Performance of "Roundup Ready" vs. Conv. (CN) Soybean Varieties in the Northern U.S., 1998
Table 2. Roundup Ready (RR) Soybean and Conv. Variety Yields by State, 1998
Bushel per Acre Yield Difference, Conv. to RR Varieties Trial Top Five Top States Mean Varieties Variety Illinois 2.0 0.0 0.0 Iowa -4.0 -4.0 -6.0 Michigan -2.0 -5.0 -8.0 Minnesota -5.0 -6.0 -5.0 Nebraska -7.0 -7.0 -6.0 Ohio -2.0 -4.0 -4.0 South Dakota -5.0 -4.0 -5.0 Wisconsin -2.0 -3.0 -3.0 Average Eight States -3.1 -4.1 -4.6 Percent Yield Drag per Acre Trial Top Five Top States Mean Varieties Variety Illinois 3.4% 0.0% 0.0% Iowa -6.6% -6.3% -9.1% Michigan -3.0% -6.8% -10.3% Minnesota -7.6% -8.2% -6.8% Nebraska -12.1% -10.8% -9.1% Ohio -3.3% -6.0% -5.8% South Dakota -10.2% -7.4% -8.9% Wisconsin -2.8% -3.5% -3.4% Average Eight States -5.3% -6.1% -6.7% Source: 'Performance of Transgenic Soybeans - Northern U.S.', E.S. Oplinger, M.J. Martinka, K.A. Schmitz. For details by State, see Table 1. Yield Performance of "Roundup Ready" vs. Conv. (CN) Soybean Varieties in the Northern U.S., 1998
2. Yield Drag Estimates by Seed Company
A second way to estimate the RR soybean yield drag is to compare the best conventional variety offered by a given seed company in a given maturity group to the best Roundup Ready variety in the same or a similar maturity group. In many cases, the only difference between the two varieties is the insertion of the genetic material that makes the Roundup Ready variety able to survive applications of Roundup herbicide. This sort of comparison comes closest to isolating the physiological impacts of the genetic transformation required to make the soybean plant tolerant to applications of glyphosate herbicide. These impacts are the cause of RR soybean yield drag. Seed and biotechnology companies are working hard now to find ways to retain tolerance to glyphosate without adversely effecting the ability of the soybean plant to grow and protect itself against pathogens, insects and other sources of stress.
We reviewed detailed state trial data from tests carried out in Southern Wisconsin (Tables 3 and 4) and in central and southern Minnesota (Tables 5 and 6). For 10 seed companies selling both RR and conventional soybean varieties in Southern Wisconsin, Table 3 reports differences in yields for the top RR variety in contrast to the top conventional, the mean RR to the mean conventional, and the lowest yielding RR compared to the lowest yielding conventional. Because of their size, Tables 3, 4, 5, 6 and 7 appear at the end of this document.
The average Roundup Ready yield drag in Southern Wisconsin across these 10 companies was -
Note how consistent these findings are compared to the percentage differences reported in Table 2. Yield comparisons across 8 states, and in Southern Wisconsin across the soybean varieties offered by 10 companies tell the same story. The RR yield drag among the top variety tested is 6-7 percent and the yield drag is consistently greater among the higher yielding varieties in contrast to average ones.
Table 4 captures just the percent differences in yields across the 10 companies based on trials conducted in Southern Wisconsin.
In central Minnesota testing (see Table 5) the RR soybean yield drag averaged 13.1 percent, or 9.3 bushels per acre across 14 comparisons. Again, the comparisons are between the highest yielding RR variety in a given maturity group to the highest yielding conventional variety in the same or very similar maturity group. In 3 of 14 cases the yield drag exceeded 20 percent. In this part of Minnesota, the highest yield drag was 27 percent and the lowest was 3 percent.
Table 6 reports the same data in southern Minnesota, a major soybean producing area where many more varieties are offered for sale. This table again matches the top RR and conventional varieties by company within a maturity group. In this region the average yield drag was much smaller, only 2.8 percent across 50 comparisons..
Differences in yields were also much more variable, ranging from a yield advantage of 18 percent to a yield drag of 15 percent. This variability suggests that in several cases genetic characteristics other than tolerance to Roundup herbicide differed across the varieties compared. Future comparative assessments of yield performance should strive to control for other genetic differences, a task that will require contacting individual seed companies to request information on the differences between varieties.
Economic Impacts of the RR Yield Drag
Table 7 projects the farm-level economic impacts of the yield drag associated with the planting of Roundup Ready soybean varieties. Note the table does not address differences in herbicide expenditures or in other facets of weed management. It simply places into perspective the two relatively easy-to-measure consequences of planting RR soybeans - the yield drag and the technology fee.
The average yield drag in bushels per acre appears in the second column, and as a percent of conventional yields in the third. The dollar value of the yield drag is based on a soybean price of $5.25 per bushel. The "Added Cost of the RR System" is the sum of the "Dollar Value of Yield Drag" plus the average $8.00 technology fee associated with the planting of an acre of RR soybeans. The last column expresses the added cost as a percent of gross income.
In four of eight states, farmers choosing to plant Roundup Ready soybeans incurred added costs equal to over 10 percent of gross income per acre. The added cost is lowest in Illinois, where it averaged about 2.3 percent of gross income per acre, and highest in Michigan, at $50.00 per acre or over 12 percent of gross income.
The added costs associated with Roundup Ready varieties are sizable compared to total seed plus herbicide expenditures, which typically run about $45.00 to $60.00 per acre. Accordingly, for some farmers the decision to grow RR soybeans may come close to doubling the cost of seed plus weed management systems when both the direct and indirect consequences of growing RR soybeans are taken into account. In Illinois and on some farms throughout the Midwest, RR varieties perform as well or even better than conventional varieties, despite the yield drag associated with all varieties. Seed and biotechnology companies are working to understand why. Answers will no doubt guide future efforts to bred improved varieties for the western and northern Cornbelt or to modify agronomic practices to better exploit the genetic yield potential in RR varieties. Clearly, additional factors need to be taken into account to come up with a complete assessment of the economic impacts of RR soybeans. The amount of money spent on herbicides per acre varies widely across the Midwest and as a function of agronomic practices and field conditions. Most data suggest that those farmers using RR soybean systems with greatest success under conditions of relatively low weed pressure can get by with one application per year and costs somewhat below the average.
But an increasing number of farmers need two applications of Roundup and one of at least one additional active ingredient. There are little or no herbicide cost-savings on such farms relative to growers planting conventional varieties and using Integrated Weed Management. A myriad of price incentives, volume discounts, product guarantees and rebates has made it very difficult to compare the actual costs of herbicide based systems and will continue to plague those making such comparisons for some time to come.
3. USDA Survey Data
Each year the U.S. Department of Agriculture (USDA) carries out an Agricultural Resource Management Study (ARMS) that includes surveys of agricultural chemical use and cropping practices in each state. On June 25, 1999 the USDA's Economic Research Service (ERS) released an analysis of 1996, 1997, and 1998 ARMS survey data entitled "Genetically Engineered Crops for Pest Management."6 This short report contains three tables that have led to much discussion and many misleading statements.
6 The report is accessible on the Internet through the ERS "Issues Center - Biotech," at http://www.econ.ag.gov/whatsnew/issues/biotech/ A link to this report and analysis of it can be found on the Ag BioTech InfoNet website at http://www.biotech-info.net/bt-transgenics.html, under "General Discussion and Opinions."
The first table in the ERS report covers the percent of acres planted to Bt-transgenic and herbicide tolerant varieties since 1996, as well as the percent of production. The data is organized by regions and aggregated to cover "All Surveyed States."
In the case of herbicide-tolerant soybeans, acres planted to Roundup Ready soybeans are combined with acres planted to sulfonylurea and imidazolinone tolerant soybeans. Accordingly, it is not possible to isolate the growth in acres planted to Roundup Ready varieties, even though it is widely accepted that most of the increase in acres planted to herbicide tolerant varieties have involved Roundup Ready seed. In all surveyed states, the percent of soybean acres planted to herbicide tolerant varieties rose from 7.4 percent in 1996 to 44.2 percent in 1998. In crop season 1999 at least 50 percent of the nation's soybeans are Roundup Ready (personal communication, Dr. Matt Liebman, Iowa State University).
Table 2 in the ERS report covers mean yields on acres planted to transgenic varieties compared to mean yields from acres planted to all other seeds. The Table reports the percent difference in mean yields on farms adopting the technologies compared to those planting all other seeds. In the case of herbicide-tolerant soybeans, the yield differences vary from an 8 percent yield drag in the Eastern Uplands region in 1998 to a 24.2 percent yield advantage in the Prairie Gateway region in 1998.
In a preface to the tables, the authors of the ERS report state several caveats about using mean yields from survey data to compare the genetic potential of different crop varieties. The authors state appropriately that conditions other than the genetic differences between Roundup Ready and conventional seed varieties might account for some or all the observed differences in yields. The authors conclude that -
"Thus, differences between mean estimates for yields and pesticide use from survey results cannot necessarily be attributed to the use of genetic engineering technology since the results are influenced by many other factors not controlled for, including irrigation, weather, soils, nutrient and pest management practices, other cropping practices, operator characteristics, pest pressures, and others." ("Caveats About Using Comparisons of Means," ERS report, "Genetically Engineered Crops for Pest Management")
We concur with the ERS analysts. The only reliable conclusions one can draw from the ERS data are that -
It is impossible to tell from the data reported by ERS what portion of the variability in yields is attributed to natural factors like soil, weather and pest pressure, or management factors, in contrast to genetic differences in the seed planted. It is likely, for example, that Roundup Ready adopters include a higher percentage of farmers who are aggressive managers covering large acreage. The bigger the farm, the greater the economic value of the flexibility inherent in Roundup Ready systems. Adopters may also choose to plant the higher-priced Roundup Ready seeds on their better land. Clearly, more detailed analyses will be required to understand more fully the differences between the agronomic practices and pest management systems on farms planting RR soybeans in contrast to conventional varieties.
4. Why Roundup Ready Varieties Sometimes Out-Yield Conventional Soybeans
No one or no seed company claims that the insertion of the genes conferring resistance to Roundup into a soybean variety increases the variety's genetic yield potential, other things being equal. Still, on some farms in some regions and in some trials, Roundup Ready soybeans out-yield otherwise similar conventional varieties. Indeed, this is one of the major reasons they have proven so popular with farmers. It is important to understand why this occurs in order to accurately project the magnitude of the Roundup Ready yield drag under truly comparable conditions.
Seed companies have worked hard to insert the RR gene and trait into their best commercial varieties for various soil types and maturity groups. There are two ways that the RR soybean version of these best commercial varieties could produce higher yields --
In most RR soybean systems growers must tolerate a certain degree of early-season weed pressure. To work, Roundup must be applied to green, growing weeds. Optimal results with RR beans require careful timing of the first application. If the application is made soon after weeds germinate, the farmer will avoid setting the crop back but might suffer some yield loss from later season weed pressure. Sometimes farmers have to apply a second application of Roundup to clean up fields, increasing the cost of the system and causing some damage to plants. The need for a second application can be avoided by delaying the first application, but this risks a certain degree of early season weed competition. Hence, RR weed management systems that rest solely on Roundup applications are likely to result in weed competition at least as great as in other well-managed systems.
In the field, farmers have come to expect a very high level of weed control, expectations reinforced by the proliferation of new chemistry, emergence of herbicide-tolerant varieties, and the marketing and performance guarantee programs offered by herbicide manufactures. Major across-the-board price reductions since 1997 are another factor leading to increased rates of application. In early 1998 Monsanto announced about a 25 percent reduction in the price of Roundup. Soon thereafter Dupont and American Cyanamid announced comparable, and in some cases even steeper reductions in the prices of their competing herbicides, as they struggled to slow the loss of market share to Roundup.
And so most experts agree that greater seed and product choice and lower herbicide prices has resulted in very aggressive and generally successful soybean weed management in recent years. If weed pressure and competition does not account for much of the observed differences in yields between RR and conventional varieties, what does?
Most farmers not using a Roundup Ready based system are applying either sulfonylurea or imidazolinone herbicides. Most university soybean trials also incorporate routine applications of these products in their side-by-side comparisons. These herbicides are known to cause some stunting and damage to soybeans early in the season, as clearly documented in the extensive performance trials carried out in 1998 by Iowa State University scientists.7 Soybeans grown in areas with long seasons tend to be better able to overcome early season soybean herbicide injury, although a variety of environmental and chemical interactions can undermine this capacity.
7 Go to the Iowa State University extension weed management homepage at http://www.weeds.iastate.edu and then to the section "1998 Weed Research Results Now Online." At each of several locations around the state, different teams of researchers carried out multiple trials assessing herbicide performance. Many studies were designed to determine optimal ways to incorporate residual pre- or post-plant herbicides into RR soybean systems. Early season injury to soybean plants from other herbicides is a recurrent focus of researchers and a growing concern among farmers.
These highly-active herbicides can also reduce soybean yields through other mechanisms such as lessening the availability of phosphorous or impairing the plant's immune response. But regardless of the mechanism - direct injury or indirect soil ecosystem or plant health impacts - it is clear that Roundup Ready systems sometimes produce higher yields because they avoid some of these problems associated with other herbicides. For this reason the RR yield drag is actually somewhat greater than suggested by current university trials. To more accurately isolate the magnitude of physiological RR yield drag under various conditions, university trials should be extended to include comparison plots where Integrated Weed Management and non-chemical systems are used.
The Devil in Weed Management Details
In general the better the growing season and the more even the crop stand, the better the chance that the soybean crop will fully recover from early season weed pressure and/or herbicide-based crop injury.
But all sorts of complicated interactions can arise between herbicides, surfactants, insecticides, application methods and timing, weather, tillage methods and soil conditions. These interactions sometimes create unforeseen problems. Wet and cool springs can delay the microbial degradation of herbicides and increase carry-over damage, while also slowing crop emergence and favoring certain hard to control weeds. Dry years can undermine the efficacy of certain herbicides and enhance the odds of a whole new set of adverse interactions. Soils with low pH levels can cause problems, as can mixing herbicides with certain soil insecticides.
These interactions are, moreover, inherently unpredictable. Their adverse consequences - and impact on the farmer's bottomline - are often unavoidable once the conditions set the interactions in motion. Farmers have so eagerly adopted Roundup Ready soybean technology, despite higher costs and evidence of a yield drag, in order to avoid these headaches and periodic losses. It is also why seed and biotechnology companies have bred IMI (imidazolinone) and STS (sulfonylurea) herbicide-tolerant soybean varieties -- to avoid soybean crop injury either from direct applications to herbicide tolerant seeds or carry-over from cornfields treated the year before with these often very persistent herbicides.
C. Impacts of Roundup Ready Soybean Systems on Pesticide Use
The benefit of planting a herbicide-tolerant variety is to gain added flexibility in when herbicides can be applied. Growers spending the extra money on herbicide tolerant seeds have made an implicit decision to largely rely on herbicides as their major weed management tool. Hence, it is no surprise that reliance on herbicides is greater in herbicide tolerant systems compared to Integrated Weed Management systems that include non-chemical practices such as rotations, cultivation, and use of cover crops.
Roundup Ready soybeans are a new technology. The first significant acreage was planted in 1997. Adoption rose dramatically in 1998 and is expected to grow further in 1999. Still relatively few acres have been planted to Roundup Ready soybeans more than two times, so it is way too early to determine the severity and impacts of weed shifts and resistance on pesticide use.
Still, the signs are not encouraging. Slipping efficacy is being reported by farmers across the Midwest, especially in control of velvetleaf and ragweed species.8 Many are reporting inadequate control of these weeds after applying Roundup Ultra at a rate of 24 ounces per acre (0.75 pounds active ingredient) but good control at 32 to 48 ounces per acre. A rate of 24 ounces worked acceptably well on these weeds in most regions of the Midwest in 1996 and 1997.
8 For example, see several posts by farmers between July 4 and July, 9 1999 on the "IPM Talk" discussion section of AgOnline, http://www.agriculture.com/scgi/Agtalk/ . Click on "IPM Talk" and review messages on the topic "reduced rate herbicides.".17
Weed shifts are occurring, perhaps rather swiftly and even the first case of RR soybean induced resistance - in waterhemp species - is either here or just around the corner. So regardless of whether RR soybeans have reduced reliance and use in 1996- 1998, it is clear that both are rising and will continue to do so until more balanced, multitactic weed management systems become the norm.
Reliance on Herbicides
Both in the near-term and longer-run it is tricky to measure changes in reliance on and use of herbicides. No single measure of herbicide use is a reliable indictor of reliance. Pounds applied per acre is misleading because some herbicides are effective at just a fraction of an ounce per acre while others require applications well over 1 pound per acre.
The biological activity and persistence of herbicides also vary greatly, as does the toxicity of herbicides. Accordingly, it is difficult to reach reliable judgements regarding changes in herbicide use without an in-depth look at just what changes have occurred, both in terms of the products applied, how often they are applied, and their rates of application. These data, coupled with measures of herbicide toxicity and persistence, make it possible to reach conclusions regarding changes in reliance over time and the risks stemming from herbicide use.
While the planting of RR soybeans will sustain a high level of reliance on herbicides, there are two ways on a given farm that Roundup Ready soybean systems can reduce herbicide "use."
First, growers might switch from "old" chemistry herbicides that are applied at 1.5 to 3.0 pounds active ingredient per acre to Roundup-based system involving maybe .75 to 1.5 pounds of Roundup per acre. Farmers using some of the new low-dose materials, or planting STS soybeans (soybeans resistant to sulfonylurea herbicides) can sometimes achieve excellent control with less than one-tenth pound of active ingredient per acre. Such reductions in herbicide use, measured by acre-treatments or pounds applied, are brought about by the switch to more biologically-active and persistent chemistry effective at lower use rates, not through reduction in reliance on herbicides.
An example -- RR soybean systems in 1998 required, on average, about 1.0 pound active ingredient per acre. This level of use is 10- to 20-times greater than the herbicide required with conventional seeds on farms using sulfonylurea herbicides, and is at least 5 times more active ingredient than several of the new low-dose herbicides like sulfentrazone. Hence, any claims that RR soybeans reduce herbicide use are incomplete without the phrase "...reduced COMPARED TO WHAT??" If the goal is minimizing herbicide use there are clearly far better - and cheaper -- ways for farmers to move forward.
Use Trending Upward
Experience in the field in 1999 suggests strongly that use of Roundup this year will rise perhaps 15 percent to 25 percent above 1998 in terms of average pounds of Roundup applied per acre. In 1998 USDA data show that the average rate per crop year for Roundup on soybeans was 0.92 pounds and there were on average 1.3 applications per acre. In 1999, use will trend upward to perhaps 1.6 applications per acre and 1.2 pounds per acre on average.
To place this level of Roundup use in perspective, in 1998 well less than 0.5 pounds of herbicide were applied to the vast majority of soybean acres not treated with Roundup. On perhaps 15 percent to 20 percent of the acres, the rate was well under 0.25 pounds. So compared to these systems, RR soybeans are heavily herbicide dependent. Moreover, because of weed shifts, resistance, price cuts and aggressive marketing, Roundup use is bound to rise sharply in the next few years, hastening the day when farmers will be forced to seek new solutions.
What comes next is the soybean farmer's $64,000.00 question. It remains to be seen whether any company or public research institution will come forward with answers that cut to the core of soybean weed management challenges. In the current economic and policy climate, this vital task might be left to growers themselves.
Table 3. Performance of Roundup Ready (RR) and Conv. Soybean Varieties by Company, 1998 Varietal Trials in Southern Wisconsin
Yield Difference: Conv. to RR . Maturity Bushels per Percent Yield Company Variety Group Yield Acre Drag Asgrow Top RR AG 2301 2.3 73 -4 -5.2% Top Conv. A 2553 2.5 77 Mean RR - - 70 -7 -9.1% Mean Conv. - - 77 Low RR AG 1901 1.9 68 -9 -11.7% Low Conv. A 2553 2.5 77 Cole Top RR Dyna-Gro 3266RR 2.5 73 -1 -1.4% Top Conv. Dyna-Gro 3252 2.5 74* Mean RR - - 69.3 -2.5 -3.5% Mean Conv. - - 71.8 Low RR Dyna-Gro 3214RR 2.1 67 1 1.5% Low Conv. Dyna-Gro 3233 2.5 66 Dairyland Top RR DSR-215/RR 2 69* -4 -5.5% Top Conv. DST2124 2 73* Mean RR - - 68.7 -3.6 -5.0% Mean Conv. - - 72.3 Low RR DSR-293/RR 2.8 68 -3 -4.2% Low Conv. DSR-277 2.8 71 Dekalb Top RR RR 2300 2.3 72 -4 -5.3% Top Conv. CX 253 2.5 76 Mean RR - - 68.1 -4.5 -6.2% Mean Conv. - - 72.7 Low RR CX 256RR 2.5 65 -4 -5.8% Low Conv. CX 230 2.3 69 Golden Harvest Top RR X 198RR 1.9 70 -6 -7.9% Top Conv. X 251 2.4 76 Mean RR - - 68 -0.7 -1.0% Mean Conv. - - 68.7 Low RR H-1238RR 2.3 67* 2 3.1% Low Conv. H-1184 1.8 65* Kaltenberg Top RR KB 215RR 2.1 68 -7 -9.3% Top Conv. KB 248 2.4 75 Mean RR - - 67.5 -5 -6.9% Mean Conv. - - 72.5 Low RR KB 249RR 2.4 67 -2 -2.9% Low Conv. KB 214 2.1 69 M/W Genetics Top RR G 2445RR 2.4 69 -8 -10.4% Top Conv. G 2380 2.3 77 Mean RR - - 68 -7 -9.3% Mean Conv. - - 75 Low RR G 2210RR 2.2 67 -8.2% -6 Low Conv. G 2711 2.7 73 Pioneer Top RR 92B71 2.7 69 0 0.0% Top Conv. 92B61 2.6 69 Mean RR - - 66 -1.3 -2.0% Mean Conv. - - 67.3 Low RR 92B01 2 64 -2 -3.0% Low Conv. 92B23 2.2 66 Stine Top RR 2091-4 2 68* -8 -10.5% Top Conv. 2499-0 2.4 76 Mean RR - - 67 -4.6 -6.4% Mean Conv. - - 71.6 Low RR 1980-4 1.9 65 -3 -4.4% Low Conv. 2500 2 68 Terra Top RR E 1980RR 1.9 72 -5 -6.5% Top Conv. E 248 2.4 77 Mean RR - - 68.7 -7.3 -9.6% Mean Conv. - - 76 Low RR TS 253RR 2.5 66 -9 -12.0% Low Conv. TS 277 2.7 75 Average Ten Companies Top RR & Conv. -4.7 -6.2% Mean RR & Conv. -4.4 -5.9% Low RR & Conv. -3.5 -4.8% *More than one variety produced the same yield. Source: Compiled by Benbrook Consulting Services, based on the 1998 Wisconsin Soybean Variety Tests, Southern Region (Table 6).
Table 4. Yield Drag Associated with the Roundup Ready (RR) and Conventional Soybean Varieties Sold by Ten Companies: 1998 Southern Wisconsin Varietal Trials
Yield Drag as Percent of Top Conventional Yield Compared to . Company Top RR Mean RR Low RR Asgrow -5.2% -9.1% -11.7% Cole -1.4% -3.5% 1.5% Dairyland -5.5% -5.0% -4.2% Dekalb -5.3% -6.2% -5.8% Golden Harvest -7.9% -1.0% 3.1% Kaltenberg -9.3% -6.9% -2.9% M/W Genetics -10.4% -9.3% -8.2% Pioneer 0.0% -2.0% -3.0% Stine -10.5% -6.4% -4.4% Terra -6.5% -9.6% -12.0% Average Ten Companies -6.2% -5.9% -4.8% Source: Compiled by Benbrook Consulting Services, based on the 1998 Wisconsin Soybean Variety Tests, Southern Region (Table 6).
Table 5. Performance of Top Yielding Roundup Ready (RR) and Conventional Soybean Varieties by Company, and Maturity Date*; 1998 Varietal Trials in Minnesota, Central Zone
Yield Difference: Conventional to RR . Maturity Bushels per Percent Yield Company Variety Date Yield Acre Drag Croplan Roundup Ready RT1557 9/17 58 -2 -3.3% Conventional L1475 9/18 60 Dairyland Roundup Ready DSR-152/RR 9/20 61 -8 -11.6% Conventional DSR-180/STS 9/20 69 Kruger Roundup Ready K-14RR 9/21 62 -8 -11.4% Conventional K-1990 9/20 70 KSC/Challenger Roundup Ready K-10RR 9/15 57 -12 -17.4% Conventional K-1414 9/17 69 Mustang Roundup Ready M-111 9/17 61 -9 -12.9% Conventional M-1160 9/18 70 Mustang Roundup Ready M-144 9/20 63 -3 -4.5% Conventional M-1167 9/19 66 Pioneer Roundup Ready 91B52 9/18 59 -4 -6.3% Conventional 9163 9/18 63 Prairie Brand Roundup Ready PB-1790RR 9/21 63 -10 -13.7% Conventional PBR-169+ 9/21 73 Ramy Roundup Ready 1555RR 9/19 55 -6 -9.8% Conventional 1525 9/19 61 Sands Roundup Ready EXP 1557RR 9/18 57 -12 -17.4% Conventional EXP 1444 9/18 69 Stine Roundup Ready 1794 9/21 57 -16 -21.9% Conventional 1680 9/21 73 Terra Roundup Ready E1181RR 9/17 62 -3 -4.6% Conventional TS107 9/17 65 Wensman Roundup Ready W2118RR 9/17 60 -17 -22.1% Conventional W3148 9/18 77 Yield King Roundup Ready K-191RR 9/22 55 -20 -26.7% Conventional K-1943+ 9/22 75 Average Roundup Ready 59.3 -9.3 -13.1% Average Conventional 68.6 Source: Compiled by Benbrook Consulting Services, based from the 'Variety Trials Results' by Crookston, Moorhead and Shelly found at http://www.extension.umn.edu/Documents/D/C/Other. *For each company, the closest match was selected by maturity date.
Table 6. Performance of Top Yielding Roundup Ready (RR) and Conventional Soybean Varieties by Company, and Maturity Date*; 1998 Varietal Trials in Minnesota, Southern Zone
Yield Difference: . Maturity Bushels per Percent Yield Company Variety Date Yield Acre Drag Asgrow Roundup Ready AG2201 9/19 64 5 8.5% Conventional A1 923 9/18 59 Asgrow Roundup Ready AG2301 9/21 60 -6 -9.1% Conventional A2247 9/21 66 Dahlman Roundup Ready 818RR 9/15 61 4 7.0% Conventional Russel 9/16 57 Dairyland Roundup Ready DSR-241/RR 9/19 64 2 3.2% Conventional DSR-180/STS 9/19 62 DeKalb Roundup Ready CX191RR 9/20 65 -3 -4.4% Conventional CX205 9/20 68 Dyna Gro Roundup Ready 3173RR 9/20 60 -3 -4.8% Conventional 3188 9/20 63 Golden Harvest Roundup Ready 165RR 9/19 67 1 1.5% Conventional H-1184 9/19 66 Golden Harvest Roundup Ready H-1207RR3 9/21 60 -3 -4.8% Conventional H-1214 9/21 63 Great Lakes Roundup Ready GL2000RR 9/24 64 -1 -1.5% Conventional GL2451 9/24 65 Kaltenberg Roundup Ready X160RR 9/20 63 -1 -1.6% Conventional KB221 9/20 64 Kaltenberg Roundup Ready KB209RR 9/22 64 -4 -5.9% Conventional KB208 9/21 68 Kaltenberg Roundup Ready KB159RR 9/23 63 -2 -3.1% Conventional KB248 9/24 65 Kruger Roundup Ready K-24APR 9/20 65 1 1.6% Conventional K-2242 9/20 64 Kruger Roundup Ready K24RR 9/21 65 0 0.0% Conventional K-2303 9/21 65 KSC/Challenger Roundup Ready K-22RR 9/19 62 -7 -10.1% Conventional K-2125 9/20 69 KSC/Challenger Roundup Ready K-191RR 9/21 65 -4 -5.8% Conventional K-2343 9/22 69 Latham Roundup Ready 406RR Brand 9/17 60 -4 -6.3% Conventional 410 Brand 9/18 64 Latham Roundup Ready EX-656RR 9/20 63 -4 -6.0% Conventional 660 Brand 9/20 67 Latham Roundup Ready EX-426RR 9/23 65 -2 -3.0% Conventional 621 Brand 9/22 67 Mustang Roundup Ready M-208 9/20 59 -3 -4.8% Conventional E-201 9/19 62 Mustang Roundup Ready M-202 9/22 64 -2 -3.0% Conventional M-2218 9/22 66 Mycogen Roundup Ready 5214 9/22 60 -8 -11.8% Conventional J-251 9/21 68 Northstar Roundup Ready 2023RR 9/18 57 -6 -9.5% Conventional 2302 9/19 63 PBR Roundup Ready PBR-1997RR 9/21 67 4 6.3% Conventional PBR-218 9/21 63 Pioneer Roundup Ready 92B05 9/15 66 10 17.9% Conventional 9163 9/15 56 Pioneer Roundup Ready 92B21 9/19 65 1 1.6% Conventional 92B23 9/19 64 Prairie Brand Roundup Ready PB-2124R 9/21 67 1 1.5% Conventional PB-197 9/21 66 Prairie Brand Roundup Ready PB-2090RR 9/24 67 1 1.5% Conventional PB-235 9/24 66 Profiseed Roundup Ready 4201 9/21 67 -2 -2.9% Conventional 2557 9/22 69 Profiseed Roundup Ready 4249 9/24 57 -10 -14.9% Conventional 2413 9/24 67 Ramy Roundup Ready 1995RR 9/20 60 -6 -9.1% Conventional 2195 9/21 66 Ramy Roundup Ready 2195RR 9/22 63 -9 -12.5% Conventional 2198 9/22 72 Ramy Roundup Ready 2085RR 9/23 65 -1 -1.5% Conventional 2550 9/23 66 Renze Roundup Ready R1909R 9/20 64 -2 -3.0% Conventional R2098 9/20 66 Renze Roundup Ready R2009R 9/23 65 0 0.0% Conventional R2297 9/22 65 Sands Roundup Ready EXP 9619RR 9/21 67 6 9.8% Conventional EXP 2027 9/20 61 Sands Roundup Ready SOI 245RR 9/23 63 -8 -11.3% Conventional EXP 2435 9/24 71 Sansgaard Roundup Ready S-2100RR 9/18 61 -6 -9.0% Conventional S-190X 9/18 67 Sansgaard Roundup Ready S-245XRR 9/20 62 1 1.6% Conventional S-199X 9/20 61 Sansgaard Roundup Ready S-233XRR 9/23 62 -10 -13.9% Conventional S-237X 9/22 72 Stine Roundup Ready 1991-4 9/21 66 1 1.5% Conventional 1970 9/21 65 Terra Roundup Ready E-1481RR 9/16 60 -4 -6.3% Conventional E1 58 9/16 64 Terra Roundup Ready E2081RR 9/20 60 -6 -9.1% Conventional TS194 9/19 66 Thompson Roundup Ready T-3178RR 9/15 58 -2 -3.3% Conventional EX7217 9/16 60 Thompson Roundup Ready T-3215RR 9/21 66 -2 -2.9% Conventional T-3222 9/21 68 Thompson Roundup Ready T-3208RR 9/23 63 -2 -3.1% Conventional EX7705 9/24 65 Wensman Roundup Ready W 2178RR 9/16 63 0 0.0% Conventional W 3148 9/15 63 Wensman Roundup Ready W2198RR 9/21 65 -6 -8.5% Conventional W3207 9/21 71 Yield King Roundup Ready K-19FRR 9/20 65 6 10.2% Conventional K-2323STS 9/21 59 Yield King Roundup Ready K-20RR 9/23 64 -6 -8.6% Conventional K2525A 9/22 70 Average Roundup Ready 63.2 -2.0 -2.8% Average Conventional 65.2 *For each company, the closest match was selected by maturity date. Source: Compiled by Benbrook Consulting Services, based from the 'Variety Trials Results' by Crookston, Moorhead and Shelly found at http://www.extension.umn.edu/Documents/D/C/Other.
Table 7. Farm Level Economic Impacts of the Roundup Ready Soybean Yield Drag in 1998: Top Yielding RR to Conventional Varieties
Average Yield Average Dollar Value Added Cost Added Cost as Number of Drag Percent of Yield of RR Percent of Companies [Bushels/Acre] Yield Drag Drag* System* Gross Income State Varietal Trial Comparisons . Minnesota: Central 14 -9.3 -13.1% $48.75 $56.75 15.8% Minnesota: Southern 50 -2.0 -2.8% $10.61 $18.61 5.4% Wisconsin: Southern 10 -4.7 -6.2% $24.68 $32.68 8.3% Oplinger Summary Data Illinois - 0 0.0% $0.00 $8.00 2.3% Iowa - -6 -9.1% $31.50 $39.50 11.4% Michigan - -8 -10.3% $42.00 $50.00 12.2% Minnesota - -5 -6.8% $26.25 $34.25 8.8% Nebraska - -6 -9.0% $31.50 $39.50 11.4% Ohio - -4 -6.0% $21.00 $29.00 8.0% South Dakota - -5 -9.0% $26.25 $34.25 11.6% Wisconsin - -3 -3.4% $15.75 $23.75 5.2% Source: Compiled by Benbrook Consulting Services from state varietal trial data and (Oplinger, et al., 1999). *Assumes a soybean price of $5.25 per bushel and an average "Technology Fee" of $8.00 per acre.
Benbrook, C, 1999. "World Food System Challenges and Opportunities: GMOs, Biodiversity, and Lessons from America's Heartland," Paper presented January 27, 1999 at the University of Illinois World Food Sustainable Agriculture Program. Accessible in Adobe Acrobat format on Ag BioTech InfoNet at http://www.biotech-info.net/costs.html#cost_returns .
Benbrook, C., Groth, E., Hansen, M., and S. Marquardt, 1996. Pest Management at the Crossroads, Consumers Union, Yonkers, New York. Drinkwater, L.E., Wagoner, P, and M. Sarrantino, 1998. "Legume-based cropping systems have reduced carbon and nitrogen losses," Nature, Vol. 396: 262-265. Hager, A., and M. McGalmery, 1998. "Corn Herbicide/Insecticide Interactions," University of Illinois Cooperative Extension. Accessible at http://spectre.ag.uiuc.edu/cespubs/pest/articles/v983f.html .
Oplinger, E. S. Oplinger, Martinka, M.J., and K. A. Schmitz, 1999. "Performance of Transgenic Soybeans in the Northern U.S.," Accessible in Adobe Acrobat format at http://www.biotech-info.net/herbicide-tolerance.html#soy .
Owen, M., 1999. "North American Developments in Herbicide Tolerant Crops," Paper presented at the 1997 British Crop Protection Conference, Brighton, England. Accessible at http://www.weeds.iastate.edu/weednews/Brighton.htm .
Tilman, D., 1998. "The greening of the green revolution," Nature, Vol. 396: 211-212. Wyse, D, 1994. "New Technologies and Approaches for Weed Management in Sustainable Agriculture Systems," Weed Technology, Vol. 8: pages 403-407.
Soybean Yield Trial Data
Minnesota: James H. Orf , Philip J. Schaus, Leland L Hardman, and Arthur S. Killam, "1998 Soybean Variety Trial Results," Minnesota Agricultural Experiment Station,. Department of Agronomy and Plant Genetics, University of Minnesota, St. Paul, MN 55108. Accessible on the Internet at http://www.extension.umn.edu/Documents/D/C/DC7349.html .
Wisconsin: Oplinger, E.S., Martinka, J.M., Schmitz, K.A., and C.R. Grau, 1999. "1998 Wisconsin Soybean Variety Tests," Department of Agronomy, College of Agricultural and Life Sciences, University of Wisconsin-Madison. Accessible in Adobe Acrobat format on the Internet at < http://www.uwex.edu/ces/soybean/ >.
If you have come to this page from an outside location click here to get back to mindfully.org