Using Crop Performance Data to Select Hybrids and Varieties

The proliferation of hybrids and varieties available to growers and the availability of data in various formats make varietal selection difficult. This report seeks to guide growers, consultants, and extension agents on how to use the most appropriate data sets from university trials in varietal selection. Six crops (barley [Hordeum vulgare L.], corn [Zeu mays L], cotton [Gossypium hirsutum L], oats [Avena sativa L.], soybean [Glycine max (L.) Merr.], and wheat [Triticum aestivum L.]) in the North Carolina Official Variety Trials were examined for turnover rate, variety × environmental interactions, and probabilities of predicting the top varieties in a subsequent year based on single and 2-yr data. Single-year turnover rates ranged from 27 to 51%. There were no variety × location interactions but there were variety × year interactions in barley and early and medium maturing cotton, and variety × location × year interactions in barley, MGVI soybean, and wheat. Based on all data examined in this study, single-year multi-location data would be appropriate to use in selecting midseason corn hybrids. Two-year multi-location data would be useful for barley, early and full-season corn, early and medium cotton, oats, MG V and VI soybean, and wheat. The protocol used in this study could be applied to other crops in other states recognizing that not all crops are tested equally and best data sets may be different for the various crops. Research QuestionUniversity-run state crop performance trials provide unbiased information for growers, consultants, extension agents, and seed company representatives to use in making varietal selection decisions. The trials include many entries from several different companies or public institutions and are located at several sites in each state. This vast data set serves to confuse the grower and agricultural professional; they have to know the most helpful data to use in their varietal selection process. Literature SummaryIn Missouri and North Dakota, 2-yr multi-location means were found to be best for choosing corn hybrids. Multi-location data were the best predictors of future performance of soybean varieties in Minnesota. Private seed companies are advising that crop performance data be explained and presented in a manner that aids in interpreting the data. Study DescriptionSix crops from the North Carolina Official Variety Trials were examined for turnover rate, significant interactions, and predictability. Ten years of data were available for barley, corn, oats, soybean, and wheat while 8 yr of data were available for cotton. The percentage of new varieties or turnover rate was determined for each year and every 2 yr. The probability of choosing high-yielding varieties or the top variety was determined from three different data sets. Applied QuestionsHow often are new varieties entered in the crop performance trials and what are the ramifications? The turnover rate ranged from 51% of the early-season corn hybrids to only 27% of the barley varieties each year. Every 2 yr, 71% of the corn hybrids are new and only 40% of the barley varieties are new. Therefore for crops like the early-season corn hybrids, choosing hybrids from 3- and 4-yr data sets would mean some of the newer and possibly higher-yielding corn hybrids on the market would be ignored. What interactions are important and what impact will they have on interpreting that data? There were no variety × location interactions for any of the six crops, suggesting that multi-location data can be used. Variety × year interactions occurred with barley and cotton, indicating a need to use a minimum of 2-yr data. The three-way interaction (variety × location × year) was important for barley, MG VI soybean, and wheat; this would suggest the use of multi-location multi-year data. For what crops can we use single-year multi-location data to choose varieties or hybrids? These data proved adequate for mid-season corn hybrids. Which crops require multi-location 2-yr data for the best predictions and how predictive are they? Barley, early- and full-season corn hybrids, cotton, oats, MG V and VI soybean, and wheat all require 2-yr multi-location data. Predictions ran from 0.54 for medium-maturing cotton varieties to 0.94 for full-season corn hybrids for choosing the high-yielding entries. RecommendationThe conclusions reached in this research pertain to North Carolina, but the principles are applicable to other states and regions. This approach should be followed whenever crop performance data are published so that growers and agricultural professionals can more efficiently select hybrids and varieties for their particular operation.