NC State Extension Publications

## Introduction

Choosing a cotton variety (or varieties) is one of the most important decisions a grower can make in a given year. Once seeds are planted, the maximum genetic potential for that field and season is determined. Beyond planting, the prevailing environment and management practices can only assist in preserving and achieving the maximum genetic yield potential for a field in a specific environment. Variety decisions encompass several considerations and criteria. Yield is usually the focus of these decisions, but also consider factors such as technology packages for weed and insect management, fiber quality characteristics, seed size and vigor, leaf hairiness, growth potential and maturity, or stormproof characteristics.

## Identify Yield-Limiting Factors

Selecting a cotton variety can be a daunting task. First, it is important to identify the most common yield-limiting factors on a field-by-field basis. Some yield-limiting factors may influence which variety will perform best in a particular field. In North Carolina, water is usually the most common yield-limiting factor. Given that most cotton acres in North Carolina are rain-fed, this factor is predominately related to rainfall patterns but can also be related to soil type and how well certain soils retain plant available water. Some varieties may only perform competitively when planted in heavier productive soils, especially in environments that receive timely rains. Other varieties may perform competitively in lighter, sandier soils that encounter drought stress. Other yield-limiting factors that could influence relative variety performance include the presence of root-knot nematodes, weed control (herbicide tolerant traits), and the ability of some fields to produce adequate stalk height necessary for acceptable yields.

## Evaluate Stability Characteristics

Second, it is essential to evaluate varieties’ stability characteristics. Any variety can perform well in an individual trial; however, it may or may not perform well in other conditions. No variety ever wins all trials, so there is no one-size-fits-all variety. Some varieties, however, consistently perform better than others. The frequency at which a variety performs well or competitively across a broad range of factors and environmental conditions indicates its ability to perform well consistently and repeatedly. All variety trials differ in some capacity. In North Carolina, summer rains are often sporadic and differ greatly from field to field and region to region. In addition, the numerous trials conducted across the state differ in terms of soil type, irrigation, planting date, management practices, etc. Therefore, varieties that can perform competitively across all of these factors indicate a high degree of stability. These more stable varieties are more likely to perform competitively and consistently in subsequent years, regardless of rainfall patterns, and across soil types and management practices. The only way to effectively evaluate variety stability is to observe replicated trial data from as many locations as possible rather than relying solely on the trial that happens to be closest to your farm. Finally, because no single variety wins all trials, it is important to identify and plant multiple varieties that are stable and consistent performers. Every variety has strengths and weaknesses; therefore, it is very risky to plant all of your acres to a single variety.

## Consider Replicated Trial Data

When selecting varieties, be sure to consider as much replicated, multi-location, and multi-year data as possible, and compare varieties under the same environmental conditions. Although convenient, field-by-field observations of variety performance on your farm are not generally recommended or at least not with much emphasis. Understanding the role of environmental influences on actual yield is crucial, and a poor understanding could result in over-emphasizing field-by-field observations and thus poor variety choices.

Any variety can achieve high yields if the environment supports high yields, while the same variety may perform poorly if the environment doesn’t support high yields. Any small variation in rainfall or soil water retention from field to field can drastically affect final yields. It is therefore an inaccurate and unfair assessment to compare a field of variety A to another spatially separated field of variety B, especially if they differ in terms of rainfall amount or timing and soil type. Some growers will not plant certain varieties that may have yielded poorly in their fields even though these varieties performed well in trials. Despite this experience, the variety in question may have been the best choice for that situation but the environment simply didn’t support high yields. In these cases, a different variety may have been an even worse choice. Others have stated that they can achieve acceptable or high yields with varieties that don’t yield as well as others in trials. This may be true as well; however, they may have achieved much higher yields with a better-performing variety and therefore missed an opportunity to be more profitable. The only effective way to evaluate variety performance is to observe replicated trial data where varieties are compared in the same environmental conditions.

## Determining the Impact of Poor Variety Decisions

So you have considered all of the aforementioned criteria for selecting varieties and you are ready to put that to practice. You may be asking yourself the following questions: (1) Are variety decisions really that important? (2) There are several companies that offer a plethora of good varieties, so I can’t go wrong, can I? And (3) More importantly, what is the cost me of making a poor variety decision? These are all excellent questions, and the answers are (1) YES!, (2) YES!, and (3) SIGNIFICANT! Many growers often disregard, or at least underestimate, the financial impact that poor variety decisions can have. This publication uses results from a wide range of replicated variety trials conducted in North Carolina to illustrate the true impact of variety decisions on profitability for North Carolina cotton producers.

This publication includes data from both the 2015 North Carolina On-Farm Cotton Variety Evaluation Program (OFCVE) and the North Carolina State University Official Cotton Variety Trials (OVT). The NC State University OVT Program has been an established and revered program for several years, with support from NC State, participating seed companies, and the NC Cotton Producers Association. The NC OFCVE Program was launched on a large scale in 2015, and is largely supported by the NC Cotton Producers Association, NC Department of Agriculture & Consumer Services, and NC State, with additional support from Cotton Incorporated, The NC Ag Foundation, and participating seed companies. Both testing programs utilize replicated trials across a broad range of environments, soil types, and geographical regions to evaluate variety performance and collectively provide producers with a complete assessment of modern varieties and a means to effectively evaluate variety stability.

The economic values illustrated in this publication reflect current cotton prices at $0.65 per pound of lint. All figures assume no discount or premium for lint quality due to the strong influence of environmental conditions on lint quality (discussed later) and the inconsistency in discount penalties or premium payments. This inconsistency is influenced by marketing avenues, buyer preferences, time of year in which cotton is sold, and how fiber quality compares to other cotton sold in a similar time frame. Current cotton prices are relatively low compared to the last several years. With that in mind, if cotton prices were to increase to$0.80 per pound, the figures illustrated in this publication, and therefore the cost of improper variety selection, would increase by 25 percent!

Table 1 illustrates several types of variety comparisons and the associated degree of error in terms of opportunity loss of potential revenue (deviation from maximum possible revenue) for both the OFCVE (On-Farm Large-Plot Variety Trials) and OVT (Small-Plot Variety Trials). The on-farm large-plot trials consist of only 10 varieties that are considered to be the best-performing and most widely adapted varieties from each seed company. Therefore, the error in variety selection should be relatively low. The OVT program consists of 40 varieties from all brands, including experimental varieties and niche varieties in addition to the varieties that appear in the on-farm program. Theoretically, the error in variety selection in these trials may be higher.

Yield Impacts

The largest degree of variety selection error is illustrated by comparing numerical yields for the highest-yielding variety in each trial to the lowest-yielding variety in each trial. Averaged over the 17 OFCVE trials, $156/acre in potential revenue was lost, which ranged from$103 to $315/acre. Averaged over the six OVT locations,$224/acre in potential revenue was lost, which ranged from $171 to$259/acre. A slightly lower degree of variety selection error is illustrated by comparing the average yield of the top 30 percent of varieties to the average yield of the bottom 70 percent of varieties in each trial. Averaged over the 17 OFCVE trials, $120/acre in potential revenue was lost, which ranged from$69 to $236/acre. Averaged over the six OVT locations,$122/acre in potential revenue was lost. This is a similar figure to that noted in the OFCVE trials; however, the range observed from the OVT locations was somewhat tighter ($99 to$154/acre). An even more conservative approach and lower degree of error is illustrated by comparing the average yield of all varieties performing above the trial average to those performing below the trial average in each trial. In this comparison, $81/acre in potential revenue was lost when averaged over the 17 OFCVE trials, ranging from$45 to $146/acre, and similarly, an average of$79/acre in potential revenue was lost in the six OVT trials, ranging from $63 to$99/acre. The most conservative and accurate approach with the lowest degree of error in variety selection is illustrated by comparing the average yield of varieties performing statistically the same as the top-yielding variety to the average yield of varieties performing statistically less than the top-yielding variety. In this scenario, an average of $66/acre in potential revenue was lost in the OFCVE trials, which ranged from$0 to $146/acre, and an average of$90/acre in potential revenue was lost in the OVT locations, which ranged from $63 to$101/acre.

 Variety Comparison (Degree of Error) Large-Plot, Replicated On-Farm Trials Small-Plot, Replicated On-Farm Trials Average Over 17 Trials Trial Where Revenue Loss Was the Least Trial Where Revenue Loss Was the Greatest Average over 6 Trials Trial Where Revenue Loss Was the Least Trial Where Revenue Loss Was the Greatest Highest Yielding Variety VS Lowest Yielding Variety -156 -103 -315 -224 -171 -259 Average of the Top 30% of Varieties VS Average of the Bottom 70% of Varieties -120 -69 -236 -122 -99 -154 Average of All Varieties Performing above Trial Average VS Average of All Varieties Performing below Trial Average -81 -45 -146 -79 -63 -99 Average of Varieties Statistically the Same as the Top Yielding Variety VS Average of Varities Statistically Dissimilar from the Top Yielding Variety -66 0 -146 -90 -63 -101

Using only the average figures from the lowest degree of error, a 1,500-acre cotton grower in North Carolina could lose $99,000 to$135,000 in potential revenue (using currently low cotton prices) by simply choosing the wrong cotton variety! For North Carolina’s cotton acreage of 380,000 acres, poor variety decisions could equate to losses of $25,080,000 to$34,200,000 in the best-case scenario! Furthermore, the impact of poor variety selection drastically increases when choices are made using methods with a larger degree of error and when cotton prices are higher.

Table 2 and Table 3 illustrate the potential for revenue losses for each individual variety in the OFCVE trials and OVT locations, respectively. These scenarios illustrate to likelihood of losses in potential revenue once a variety decision is made. The purpose of this publication is not to target specific, commonly-planted varieties because varieties in each testing program continually change from year to year; therefore, the varieties are simply listed by letter in the OFCVE trials and by number in OVT. Varieties are ranked in descending order according to their average yield (lbs/acre) across the 17 OFCVE trials and the six OVT locations in each table (columns 1 and 2). For each of these tables, figures are reported by comparing the top-yielding variety in each trial to those that performed statistically less than the top-yielding variety (method with the lowest degree of error in variety selection). The third column in each table illustrates the average loss in potential revenue for each variety across all trials, including the trials where the variety performed statistically the same as the top yielder (in these locations, revenue losses are $0/acre). The fourth column illustrates the percentage of trials in which a loss in potential revenue occurred (trials where the variety yielded statistically less than the top-yielding variety), and the fifth column lists the average loss in potential revenue only in the trials where potential revenue was lost. The sixth and last column illustrates the worst-case scenario for each variety or the trial where potential revenue loss was greatest. The purpose of these tables is not only to illustrate the vast range in potential economic losses among our commonly planted varieties, but also to reiterate the point that no variety wins all trials nor is a single variety going to perform the best in every situation. When planting a single variety across all environments, expect losses in potential revenue. As seen in Table 2, average losses in potential revenue among the varieties in the OFCVE trials ranged from$13 to $85/acre. The frequency in which varieties resulted in potential revenue losses ranged from 18 to 65 percent of trials; therefore, some varieties clearly resulted in potential revenue losses much more frequently than others. When evaluating the average loss in potential revenue in only the trials where a statistical loss occurred, losses ranged from$55 to $139/acre, illustrating a very wide range in economic impact and how frequently a variety may result in less than maximum revenue. It’s important to remember that the varieties that were planted in the OFCVE trials were considered to be the best-performing varieties from each brand. In the worst-case scenario for each variety, losses in potential revenue ranged from$71 to $315/acre.  Variety Average Yield across 17 On-Farm Trials (lbs/A) Average Deviation from Maximum Possible Revenue ($/A) Percentage of Trials in Which Maximum Possible Revenue Was Not Achieved (%) Average Deviation from Maximum Possible Revenue in Trials Where Loss Occurred ($/A) Greatest Deviation from Maximum Possible Revenue ($/A) A 988 -19 18 -107 -136 B 948 -28 35 -81 -146 C 948 -13 24 -55 -71 D 925 -47 47 -100 -160 E 906 -55 53 -104 -213 F 900 -55 47 -116 -201 G 897 -59 59 -100 -249 H 880 -71 65 -110 -204 I 851 -82 59 -139 -225 J 847 -85 65 -132 -315

Table 3 illustrates the same aforementioned parameters for the six OVT locations. Remember that the OVT program includes a much larger number of varieties from essentially every brand, and nearly all are commonly planted to various degrees in North Carolina. Average losses in potential revenue for these varieties ranged from $29 to$159/acre, a somewhat wider range than that observed in the OFCVE trials. The frequency with which varieties resulted in a potential revenue loss ranged from 33 to 100 percent of trials, suggesting that some varieties will always result in some degree of loss in potential revenue. In the trials where a statistical loss in potential revenue occurred, average losses ranged from $63 to$179/acre, which further demonstrates a wide range in impact in addition to frequency of loss. In the worst-case scenario for each variety, losses in potential revenue ranged from $85 to$259/acre.

 Variety Average Yield across 6 OVT Locations (lbs/A) Average Deviation from Maximum Possible Revenue ($/A) Percentage of Trials in Which Maximum Possible Revenue was not Achieved (%) Average Deviation from Maximum Possible Revenue in Trials Where Loss Occurred ($/A) Greatest Deviation from Maximum Possible Revenue ($/A) 1 965 -34 33 -102 -141 2 960 -32 50 -63 -85 3 959 -29 33 -88 -95 4 957 -40 33 -119 -140 5 953 -51 67 -76 -90 6 945 -47 50 -94 -105 7 942 -67 67 -101 -145 8 940 -42 50 -84 -116 9 934 -77 67 -116 -216 10 934 -77 67 -115 -177 11 931 -73 67 -110 -125 12 931 -56 50 -112 -148 13 926 -82 83 -98 -138 14 921 -81 67 -121 -146 15 916 -75 67 -112 -125 16 916 -91 83 -109 -136 17 914 -94 83 -113 -203 18 912 -89 83 -106 -133 19 905 -81 50 -162 -172 20 905 -92 67 -138 -192 21 904 -76 67 -113 -155 22 902 -88 83 -106 -134 23 900 -85 67 -128 -163 24 898 -92 83 -111 -172 25 896 -96 83 -116 -127 26 894 -95 83 -114 -183 27 891 -110 100 -110 -140 28 890 -93 83 -112 -164 29 884 -114 83 -136 -174 30 880 -104 83 -125 -208 31 871 -113 83 -135 -259 32 868 -115 83 -139 -247 33 863 -114 83 -137 -190 34 858 -114 83 -137 -207 35 857 -125 83 -150 -221 36 843 -134 83 -161 -229 37 840 -143 100 -143 -199 38 833 -147 100 -147 -188 39 815 -159 100 -159 -258 40 813 -150 83 -179 -220 As mentioned previously, there is no one-size-fits-all variety that will be the highest-yielding variety in every situation. Some varieties do perform well more consistently than others, and both the frequency and magnitude of losses in potential revenue differ vastly among our most popular varieties. However, it is always recommended that growers plant multiple top-performing varieties and customize variety selection to match various environments that suit each variety. It is never advised that a grower plant a single variety across all acres. Assuming the best variety is chosen using a method with the least degree of error, and using the best case scenario illustrated in Tables 2 and 3, a 1,500-acre North Carolina cotton grower who plants a single variety across all acres could expect to lose$19,500 to $43,500 in potential revenue on average, and to lose potential revenue on 18 to 33 percent of their acres if they happen to choose the best variety on average. This equates to losses in potential revenue ranging from$4,940,000 to \$11,020,000 statewide.

Fiber Quality Impacts

In addition to yield, fiber quality grades can influence profitability and potential revenue gains or losses. It is important to note, however, that while the majority of revenue is determined by yield, it can be slightly increased with premium payments or decreased with discount penalties, or result in no adjustment at all, depending on fiber quality. The magnitude and frequency of discounts or premiums depends on many factors, including variations in marketing avenues, buyer preferences or requirements, time of year in which cotton is sold, and how your cotton may compare to other cotton being sold in a similar time frame. Therefore, frequency and magnitude of discounts or premiums are widely inconsistent and, as such, are not included in the figures discussed in this publication.

In addition, fiber quality is influenced largely by the environment and somewhat by genetics. The degree of environmental influence versus genetic influence varies for each fiber quality parameter. With that said, there is an obvious genetic influence on some of the more important fiber quality parameters such as micronaire and fiber length. We encourage growers to evaluate fiber quality as standard practice and attempt to plant varieties that are both high-yielding and high-quality. Because discounts and premiums are inconsistent, growers should strive for the highest yields while avoiding potential discounts for poor fiber quality. Although the environment greatly influences fiber quality, some varieties are more likely than others to result in poor quality. Table 4 illustrates values for both micronaire and fiber length, as well as the percent of trials in which a variety resulted in discount high micronaire, discount short fiber length, and both high micronaire and short fiber length for all varieties in the OFCVE trials. The varieties listed in Table 4 correlate to those in Table 2. It is important to understand that additional penalties or losses in potential revenue could occur by planting a variety with poor quality tendencies in an environment that promotes poor quality.

 Variety Average Micronaire (microns) Incidence of Discount High Micronaire (% of trials) Average Fiber Length (inches) Incidence of Discount Short Fiber Length (% of trails) Incidence of Both High Micronaire and Short Fiber Length (% of trials) C 4.97 59 1.11 12 12 G 4.89 47 1.09 18 18 B 4.87 53 1.11 6 6 A 4.85 35 1.08 47 29 J 4.75 35 1.08 35 18 D 4.70 35 1.07 41 24 E 4.68 24 1.13 6 6 I 4.64 29 1.14 0 0 H 4.63 29 1.07 41 24 F 4.34 12 1.18 0 0

## Conclusions

Several valuable resources are available to help you make the best variety decisions possible. Your county Cooperative Extension agent and NC State Extension cotton specialists are here to help with these and other decisions. In addition, the North Carolina Cotton Variety Performance Calculator is a free online resource that allows users to customize variety comparisons using replicated data from the North Carolina On-Farm Cotton Variety Evaluation Program and NC State University Official Variety Trials for Cotton. Users can compare varieties across a customized and wide range of environments based on many layers of criteria, including year, multi-years, trial type, yield environment, and location. The NC Cotton Variety Performance Calculator can be found on the NC State University Cotton Portal website.

## Acknowledgements

The authors would like to acknowledge and thank the NC Cotton Producers Association, NC State University, the NC Department of Agriculture & Consumer Services, Cotton Incorporated, the NC Ag Foundation, and participating seed companies for their financial support and in-kind contributions to the variety testing programs mentioned in this publication. Also, the authors would like to recognize and thank all of the many people who contribute to our variety testing programs on an annual basis, including NC State University Cooperative Extension agents, NCDA&CS Research Station staff, consultants, and especially the cooperating growers.

# Authors

Extension Cotton Specialist
Crop and Soil Sciences
Professor of Crop Science & Extension Cotton Specialist
Crop and Soil Sciences