Bill Cox1, Eric Sandsted1, and Gary Bergstrom2
1Soil and Crop Sciences Section, 2Plant Pathology and Plant-Microbe Biology Section – School of Integrated Plant Science, Cornell University
We have conducted numerous studies over the last 10 years on soybean seeding rates. In almost all instances, we rarely observed a response to rates above 150,000 seeds/acre (http://scs.cals.cornell.edu/sites/scs.cals.cornell.edu/files/shared/documents/wcu/Vol16No22006Mar-Apr.pdf; http://scs.cals.cornell.edu/sites/scs.cals.cornell.edu/files/shared/documents/wcu/WCUVol18No3_May-Jun2008-2.pdf; http://scs.cals.cornell.edu/sites/scs.cals.cornell.edu/files/shared/documents/wcu/wcu19-1.pdf; http://scs.cals.cornell.edu/sites/scs.cals.cornell.edu/files/shared/documents/wcu/WCU21-2.pdf; http://scs.cals.cornell.edu/sites/scs.cals.cornell.edu/files/shared/documents/wcu/WCU_vol22_no2.pdf). In 2015, we examined the response of soybean under high input management (~200,000 seeds/acre with a fungicide/insecticide seed treatment as well as a foliar fungicide, Priaxor, applied at 4 fluid oz. /acre on 7/31 at the R3 stage) compared with recommended input management (150,000 seeds/acre with a fungicide/insecticide seed treatment). This was part of a larger study comparing the response of all the crops in the corn-soybean-wheat/red clover rotation under conventional and organic cropping systems with recommended or high input management. We described this study in detail in a news article in the last What’s Cropping Up? issue of 2015 (http://blogs.cornell.edu/whatscroppingup/2015/11/09/soybean-yield-under-conventional-and-organic-cropping-systems-with-recommended-and-high-inputs-during-the-transition-year-to-organic/). In that article, we reported that soybean under high input management in the conventional cropping system yielded 48.6 bushels/acre, when averaged across three previous 2014 crops, compared with 44.7 bushels/acre under recommended management inputs. We did note in the previous article that despite the 9% yield increase, partial profit was not significantly different between the two treatments because the added input costs for seed and fungicide as well as the fungicide application cost offset the 3.9 bushel/acre increase for $8.50 soybeans.
Still, why did we get a yield response to high input management in the conventional cropping system? Climatic conditions were exceedingly dry from the day of fungicide application through harvest (1.4 inches of precipitation in August and 1.35 inches from September 1-20). Consequently, visually discernible disease symptoms were absent on the visible portion of soybean foliage from the R3-R8 stage. Nevertheless, because we almost never observe a response to a seeding rate of 200,000 seeds/acre compared with 150,000 seeds/acre, we speculated that perhaps the fungicide and not the higher seeding rate provided the 9% yield increase. In addition to disease control, there is some speculation that a fungicide application may improve overall soybean health, independent of disease presence.
Yield component analyses help explain why there was a yield response to high input management of soybean in the conventional cropping system in 2015. In each soybean plot (planted in 15-inch rows), we subsampled the four center rows in two 1 meter lengths (1.52 m2 area) about 5 days before the 9/23 soybean harvest of the entire plot with a plot combine. We hand-harvested all the plants in the sub-sampled area and dried them, pulled off the pods and counted them, ran the pods through a stationary thrasher, counted the seeds that were thrashed, and weighed all the seeds (~3500-4500 seeds/sample or ~7000-9000 seeds/plot).
When averaged across the three previous crops of 2014, plant populations in the sub-sampled areas a few days before harvest averaged 41.7 plants/m2 (~169,000 plants/acre) in the high input management treatment compared with 31.1 plants/m2 (126,000 plants/acre) in the recommended management treatment (Fig.1). Our sub-sampled population was slightly lower than population measurements at the V2 stage in the high input management treatment (~174,000 plants/acre, http://blogs.cornell.edu/whatscroppingup/2015/09/16/emergence-early-v2-stage-plant-populations-and-weed-densities-r4-in-soybeans-under-conventional-and-organic-cropping-systems/). In contrast, our sub-sampled population in the recommended management input treatment was slightly higher than population measurements at the V2 stage (~122,000 plants/acre). Soybean populations of the subsampled regions immediately before harvest, however, are only 2.9 to 3.2% different than at the V2 stage so we believe that our sub-sampled measurements represent the entire plot extremely well.
When averaged across the three previous crops, soybean averaged fewer pods/plant in the high input management treatment (24.5 pods/plant in high input management compared with 35.1 in recommended input management) in the conventional cropping system (Fig.2). Typically, soybeans at higher compared with lower seeding rates produce fewer pods/plant because of less intra-plant competition for light, water, and nutrients. We have seen the same degree of reduction in pods/plant at 200,000 seeds/acre compared with 150,000 seeds/acre repeatedly in previous experiments at the Aurora Research Farm (file: aj-103-1-123.pdf; file: aj-102-4-1238.pdf).
When averaged across the three previous crops, soybean averaged more seeds/pod in the high input management treatment (2.23 seeds/pods compared with 2.09 seeds/pod in recommended input management) in the conventional cropping system (Fig.3). Furthermore, soybean averaged greater seed weight in the high input management treatment (150.4 mg compared with 144.3 mg in the recommended input management treatment) in the conventional cropping system (Fig.4). In the two previously cited studies, we did not see a seeding rate effect on seeds/pod or seed weight, which is consistent with other studies in the USA. Despite virtually zero disease pressure in 2015 because of the very dry August and first half of September conditions, could the fungicide application have resulted in more seeds/pod and seed weight because of improved plant health? Certainly an increase in seeds/pod and seed weight are the two yield components most likely to be affected by a fungicide application at the R3 stage. We did not, however, observe delayed senescence in the high input management compared with the recommended input management treatment. (It is possible that the higher plant populations offset a potential fungicide effect on delayed senescence.)
If you multiply out the yield components (plants/m2 x pods/plant x seeds/pods x seed weight), the sub-sample yields averaged 51.1 bushels/acre in the high input management compared with 49.0 bushels/acre in the recommended input management treatments. The sub-sample yields were ~5-10% greater than the actual plot yields, and the difference in yield between treatments was only 4.3% instead of the 9% difference in plot yields. Nevertheless, we feel that the sub-sampled data provide us with insight on how soybean responded to high input management (fewer pods/plant with higher seeding rates and more seeds/pod and seed weight perhaps because of the fungicide application?).
We will continue this study for two additional years to see if there is indeed a response to fungicide application in the presence of high seeding rates. Currently, it is pure speculation that a fungicide application at the R3 stage resulted in more seeds/pod and greater seed weight in soybean. Nevertheless, an increase in seeds/pod and seed weight would be the two most likely yield components that an R3 fungicide application would affect. As noted previously, the 9% yield advantage did not result in an increase in partial profit. Furthermore, we avoided the harvest rows when applying the fungicide application at the R3 stage. In a production field, an R3 fungicide application would probably result in some yield reduction, associated with mechanical damage of the crop, especially with spray booms less than 90-120 feet in width, reducing the magnitude of the potential yield response. We are looking forward to 2 more years of research on this topic.