Understanding small grain yield trials
The University of Minnesota and North Dakota State University conduct yield trials to evaluate new and existing small grain varieties and compare performance and agronomic characteristics. These annual trials develop performance data over a number of locations and years, which are also referred to as environments.
Factors affecting a variety’s performance
Not only do varieties have different yields in different environments, but yield relative to other varieties also may vary depending on the environment. This yield difference is defined as a genotype by environment interaction.
Understanding how genotypes perform in different environments is important for selecting a variety that’ll perform best for you year in and year out. A variety’s performance in different environments can be mathematically described as:
Performance = Variety + Variety by environmental interaction + Environment
The genotype (genetic makeup) of a grain—such as wheat, barley or oats—variety is fixed, meaning it doesn’t change from year to year. This is why performance differences between locations and over years are a function of the environment and the variety by environment interaction.
The variety by environment interaction can be best explained graphically, as in Figure 1. In plot A, both varieties respond equally well to increased fertility. In plot B, variety 1 responds with larger yield increments to the same increase in fertility compared with variety 2. In other words, variety 1 takes better advantage of the additional fertility.
In plot C, variety 2 initially outperforms variety 1 until the fertility reaches a certain level. Once past that level, variety 1 once again is the best yielder. The performance of the varieties in plot C results in a rank change in the varieties depending on the level of soil fertility, which can affect variety selection.
The type of responses in plot C are common. Although this example uses fertility, different factors that influence grain yield, such as temperature, moisture or presence of diseases, can create similar changes in relative performance.
Testing and selecting
The existence of variety by environment interaction requires replication over a number of locations and years. This way, the total number of trials represents the greatest number of environments and conditions.
If this type of testing can be achieved, the variety rankings will likely be similar from one year to the next, and yield trial results become a useful tool to select varieties. Also, because of variety by environment interaction, absolute yield is less useful in variety trial data, and relative performance is more meaningful.
The third component in the formula is the environment and error. This portion of the performance difference can’t be explained by the variety itself or the variety by environment interaction.
Soil conditions are never uniform throughout a trial. Replication and random placement of each variety within each trial helps provide a fairer test for the variety and allow researchers to estimate the error in the trial.
Least significant difference (LSD)
The error estimate is used to calculate the least significant difference (LSD) when comparing yields. The LSD is a statistical method to determine if the observed yield differences between two or more varieties is due to varietal differences or interactions with other variables such as a difference in soil fertility within a trial.
If the yield difference between two varieties equals or exceeds the LSD value, the higher-yielding variety is considered superior in yield. If the difference was less, the yield difference may have been due to environmental error, such as soil variability, rather than genetic differences. We’d be unable to distinguish which is the better of the two.
An LSD at a 0.05 significance level indicates that, with 95 percent confidence, the observed difference is indeed a true difference in performance. Lowering this confidence level allows more varieties to be different from each other, but increases the chances of drawing false conclusions.
Reviewed in 2018