David I. Bransby, Inacio C. Maposse, Kathryne M. Glass, and Paul L. Mask

Selecting the best corn or wheat variety to plant next season may sound like a simple task, especially with all the crop variety reports published annually by the AAES. But after carefully studying one of these reports, the reader could be more confused than informed: old varieties; new varieties; one-, two-, and three-year averages; LSDs and CVs. And in every column the ranking of the varieties is different, so how can sense be made of all these numbers, and how can a choice be made with confidence? This article attempts to answer these and other questions, and shows that the approach to select the best variety may be different for wheat and corn.

The AAES crop variety testing program is large, including corn, cotton, peanuts, soybeans, small grains (wheat, rye, oats, barley, and triticale) for both grain and forage, and ryegrass, spread over 11 locations. In total, it involves planting and harvesting 7,000 plots in 87 different experiments each year. Results are reported in seven annual reports, which usually total about 160 pages and which constitute some of the most widely used publications of the AAES. However, it is not certain that this large volume of data, generated over more than 30 years, is being used to the best advantage.

Perhaps the first question to be addressed is whether the choice of a crop variety is really important. The answer to this is easily found in any crop variety report. For example, in 1997, production from the lowest yielding wheat variety averaged across 10 locations in Alabama was 30.9 bushels per acre, while that of the highest yielding variety was 61.8, exactly double. At a price of $4 per bushel, this amounts to a difference of $123.60 per acre in revenue. Even when compared to the average yield of all varieties tested (46.4 bushels per acre), the best variety produced 15.4 bushels per acre more, and provided $61.60 more revenue per acre.

Corresponding results for corn in 1997 were 100.3 and 146.7 bushels per acre for the lowest and highest yielding varieties, respectively, with an average of 122.8 bushels per acre. When compared to the average, the highest yielding variety produced 19% more, and provided $67.25 more revenue per acre if corn was priced at $2.50 per bushel. These figures alone demonstrate clearly that choice of a crop variety can strongly affect profitability, and is therefore very important.

Crop variety reports are intended to assist in selecting suitable varieties. However, data in these reports can be used in different ways to do this, and different varieties may be selected, depending on what selection criteria are used. For example if a farm is near the Gulf Coast Substation in Fairhope the latest one-year yield data or the three-year average data from that station could be used to select a variety. Alternatively, variety selection could be based on the one- or three-year yield data for the southern region of the state, which is made up of average yields from Fairhope, Monroeville, Brewton, and Headland. However, each selection criterion could suggest that a different variety is best.

Clearly, no procedure for selecting a variety is without weakness. Use of three-year averages, which is perhaps the most common, means that new varieties will not be used on farms until they have been tested for three years. Especially for superior varieties, this is an unreasonable delay. On the other hand, results from just one year may be unreliable because of specific weather conditions of that particular year. Similarly, results from only one test site may be too specific, while regional averages may not be specific enough.

To evaluate different options, data from Alabama corn and wheat variety tests were analyzed over several years and locations. For corn, varieties were selected and yields checked for 1992 to 1996 across six locations. Varieties were selected based on the following criteria: (a) the three-year regional average, (b) the previous year regional average, (c) the three-year location average, and (d) the previous year location results. Yields of the selected varieties were then recorded for each year. For example, data from the 1989, 1990, and 1991 records were used to select a variety for each of the four criteria. Yields of the selected varieties in 1992 were then recorded for each criterion. This process was repeated for each of the six locations over five years, and for wheat it was repeated for five locations over seven years.

Results indicate that different selection criteria work best for wheat and corn. Using the results from the previous year only at a given location resulted in selection of corn varieties that yielded 4.8 bushes per acre more than if varieties had been selected from the three-year regional average (see table). If corn is priced at $2.50 per bushel this amounts to a difference of $12 per acre, and $3.36 million over the 280,000 acres of corn planted in the state. For wheat, varieties selected on the basis of the three-year location average yielded 3.8 bushels per acre more than those selected from the results of the previous year only at that location. If wheat is priced at $4 per bushel, this amounts to a difference of $15.20 per acre.

In summary, use of test results from the previous year only at a given location was the best variety selection criterion for corn, while the three-year location average was best for wheat. This difference could be due to a higher variety turnover rate for corn than for wheat.