How Kernel Weight Varies by Hybrid in Iowa

Written by Ryan Van Roekel, Ph.D.; Dennis Holland; Alex Woodall; Bill Long; Matt Vandehaar; Nate LeVan and Jason Kienast, all with Pioneer Agronomy Sciences

Background and Rationale

• Corn grain yield can be estimated in-field based on estimates of yield components: ears per acre, kernels per ear, and kernel weight.
• The first two components are relatively straightforward to estimate – conducting several stand counts of 1/1000th of an acre can provide an estimate of ears per acre and kernel counts can be used to estimate kernels per ear.
• Furthermore, new technology has greatly improved the speed and accuracy of estimating the first two of the components:
  • UAV imagery powered by Drone Deploy can provide field-wide stand counts.
  • The Yield Estimator tool in the Granular Insights app will quickly count kernels per ear.
  • The Vegetation Index from satellite imagery in Granular Insights can be used to guide sampling according to field variability to get a better estimate of whole-field yield.
• However, estimating the third yield component – kernel weight – remains challenging.
• A common practice is to assume 90,000 kernels/bushel, but this practice often underestimates yield and does not consider differences among hybrids or environments.
• While work is underway to develop a more reliable way to estimate kernel weights, research was undertaken to characterize common hybrid families in local plots to provide an estimate as to how genetics influence kernels weights to provide more accurate yield estimates.
• Additionally, knowing a hybrid’s expected kernel weight can help with understanding the yield impact of late-season management or environmental issues that may prevent a hybrid from reaching its normal kernel weight.

Study Description

• Kernel weight data was collected from a selection of plots across Iowa from 2016-2021.
• Kernel weights for each hybrid at a location were measured in one of two ways:
  • A subsample of 100 random kernels, or more, were weighed and corrected to 15% moisture.
  • Multiple stand, ear, and kernel counts were performed prior to harvest to provide a reasonably accurate estimate of ears per acre and kernels per ear. This data was divided by the hybrid’s yield at 15% to determine kernels per bushel.
• Both methods have limitations, but hybrid trends were consistent, and the datasets were combined to increase the number of locations.
• A location average kernel weight was calculated from the average of all hybrids at each plot location.
• To account for environmental differences between locations, kernel weight for each hybrid within a location was calculated as a percentage of the location average. Those percentages were then averaged by hybrid family over all plot locations, as shown in Table 1.
• The standardized kernels per bushel in Table 1 were calculated as 80,000 kernels/bu divided by the average kernel weight percentage to provide a reasonable estimate for kernels/bu by hybrid family. This value is not the actual mean of the observed kernels/bu because the dataset is unbalanced for locations between hybrids. As such caution should be used with these results.

Figure 1. Representative kernels from the tip, middle, and butt of an ear from hybrid families with above-average (P1197) and below-average (P1082) kernel weight in 2019. (Photo courtesy of Bill Long.)

Table 1. Kernel weight as a percentage and standardized kernels/bu by hybrid family.

¹Calculated as hybrid kernels per bushel compared to the location average kernels per bushel, then averaged over all locations.
²Calculated as the kernel weight percentage applied to a “normal” value of 80,000 kernels per bushel, rounded to the nearest 500.

Results

• Kernel weight (kernels/bu) was found to vary widely by hybrid and location.
• The grand mean of all kernel weight observations was 82,124 kernels/bu but ranged from 52,192 to 136,518 kernels/bu. Grain yield averaged 217.3 bu/acre with a range from 116.2 to 297.3 bu/acre.
• Individual hybrids also had a wide range in kernel weights between locations. For example, the P1197 family ranged from a high of 54,656 kernels/bu down to 115,749 kernels/bu. However, across all locations, its kernel weight averaged 105.7% of the location average.

Discussion

• With the wide variation in observed kernels weights between hybrids and locations, it is important to exercise caution when using the standardized kernels/bu shown in Table 1.
  • Environmental and management factors can and will greatly influence a hybrid’s ability to maintain or extend its grain fill and express its full kernel weight potential.
    • For example, the location average kernel weight in 2020 was 85,962 kernels per bushel compared to 76,950 in 2019.
  • Often issues like drought, disease pressure, or nitrogen deficiencies can hinder late season plant health and limit a hybrid’s grain fill period and resulting kernel weight.
• It is important to note that high kernel weights are not required for high yields.
  • P1366 is an example of a hybrid family with below average kernel weight that is capable of very high yields (up to 297 bu/acre in this study).
  • P1366 tends to achieve high yields through kernel number (more rows around and/or ear length) vs hybrid families like P1197, which tends to have more average kernel numbers but high kernel weights.
• Also note that kernel weight is not correlated with test weight. Test weight is the weight of a volumetric bushel, while kernel weight is a measure of how many kernels are in a 56 lb bushel.
  • An example of this distinction is the P1093 hybrid family, which has very high test weight with excellent grain quality but its high-density kernels tend to be smaller in size and thus weigh less per kernel.
  • Contrarily, the P1197 hybrid family tends to have less dense, lower test weight grain but very large kernels that result in high kernel weights.
  • The P1213 hybrid family tends to have both high test weight and high kernel weight.
• When estimating yields, it is best to stick with an average kernel weight estimate of 80,000 kernels/bu for most hybrids.
  • Consider using a lower kernels/bu (i.e. 75,000) for hybrid families like P0306, P1197 & P1587 and higher kernels/bu (i.e. 90,000) for hybrid families like P9492 & P1093.
  • If late-season growing conditions are excellent, using a factor of 70,000 kernels/bu may be more appropriate.
  • Conversely, if late-season conditions are poor, a factor of 90,000 kernels/bu might be more accurate.
  • Be sure to get multiple, accurate estimates of kernels/ear and ears/acre to avoid overestimating yield.

Conclusions

• Kernel weight is a key component of corn grain yield that varies greatly by hybrid and environment.
• Having an idea of a hybrid’s normal kernel weight can be useful for more accurate yield estimates.
• This knowledge also helps provide an understanding of how a hybrid makes its yield (kernel number vs kernel weight), which can be useful when making management decisions or when diagnosing yield results that differ from expectations.