|
Blending two or more varieties of hard red spring wheat in the same field continues to be of interest. The reasons producers express an interest are primarily twofold. First, growers have a desire to combine high yield with high grain protein to avoid discounts. Secondly, growers want to reduce the risk of lodging by
combining lodging resistant varieties with varieties with poorer straw strength.
The idea of purposely blendingdifferent varieties of wheat is more than 50 years old and was first proposed
and tested to reduce the impact of stem and leaf rust. The rationale was as simple as it was elegant; by putting multiple resistance genes against the different races of rust in a field and across the landscape, a
leaf rust epidemic would be slower to develop. Since the epidemic would develop slower, it would take even longer for the pathogen to become more virulent. An epidemic will slow because of two factors. First, there
are decreased odds that spores will land on a susceptible genotype. Two, the presence of susceptible genotypes decreases the selection pressure for virulence (Figure 1). Ultimately, this strategy would allow
producers to limit the losses from leaf rust and breeders would be able to maintain the effectiveness of the different resistance genes longer. The concept did get some traction, but was quickly phased out in favor
of single resistance genes and monocultures as effective fungicides became available and the extra time and effort to develop and market these blends no longer could be justified.
More recently this concept of stability has not only been applied to traits like disease resistances but
also to yield and grain quality. Bowden et al. (2001) conducted a number of studies across Kansas in which a number of hard red winter wheat cultivars were blended in 2 and 3-way blends and evaluated for grain yield
potential and stability. They found that a number of the blends tested showed a slight yield advantage and that the stability of the blend was slightly higher than that of the individual varieties that were part of
the blend. Cowden et al. (2007) found that in North Carolina, blends out yielded the pure varieties by an average of 2.3 bu/A or 3.2%. The North Carolina group also found that blends were either beneficial or
neutral for incidence and severity of diseases.
Based on the literature, Doug Holen and I initiated a small blending study as part of the Red River On-Farm
Yield Trials. The objective of the study was to compare the performance of variety blends to the pure stands of the individual hard red spring wheat varieties. The traits of interest were grain yield, grain protein
content, straw strength, lodging, and incidence and severity of leaf rust.
Instead of just adding a number of 1:1 mixes, we opted to add also 1:3 and 3:1 blend of each combination of
two varieties. The additional two blends allowed us to analyze the data using linear regression and determine whether the response of the blend at the different ratios was linear. For example, does blending a 14%
protein variety with a 12% protein variety equal 13% when you harvest it? Was the observed response equal to the calculated value based on the recorded values of the pure stands of the individual varieties or was
the response non-linear (Did blending a high protein with a low protein still end up with low protein?). A non-linear response would indicate that the blend was advantageous or detrimental compared to the
monocultureof the individual varieties that made up the blend.
To limit confounding effects, we chose varieties with relative similar heading dates and which contrasted
significantly in only one dimension. For grain yield and grain protein we blended Faller and Glenn, while Bigg Red and RB07 were blended to measure the effect blending on incidence and severity of leaf rust. To
measure the effect of blending on straw strength, Rush and Traverse were blended.
For each pair of varieties, a 3:1, 1:1, and 1:3 blend was added to the existing variety trial that included
all the individual varieties. A randomized complete block with 3 replications was used for experimental design. Leaf rust and leaf disease incidence and severity were rated on 10 randomly selected flag and
penultimate leafs approximately 21 days after heading as the crop approached the soft dough stage. Lodging notes were taken on a plot basis just prior to harvest, while grain yield was determined by harvesting the
whole plot, weighed and expressed as bu/A (corrected to 13.5% grain moisture).
Initial analysis of the data indicated that for grain yield, the response to blending was linear for all
three blends (Figure 2 & 3). This means that the response to blending was additive. The response to blending for grain protein, test weight, and lodging was linear as well for all three blends (data not shown)
.The data analysis of leaf rust was not completed yet. Neither does the data provide any insight to whether
blends are more stable across locations and/or years as this was a single location. We plan to repeat the experiment again in 2009 at a number of the locations of the Red River On-Farm Yield Trials to test both mean
performance and stability of performance across environments.
For now, the bottom line is simple -- the initial results of this study suggest that there is little-to-no
incentive to blend varieties to create a yield advantage or improve grain protein that can not be accomplished by growing two adjacent pure stands. Blending the varieties in the field does not accomplish anymore than blending the varieties in your bin, postharvest
|
