1995 MISSISSIPPI COTTON VARIETY TRIALS

ACKNOWLEDGMENT

Frank Mitchener, Sumner

Keys Arnold, Tunica

Mark and Louis Fratesi, Elizabeth

Participating Companies and Varieties Entered

Terra International, Inc. P. O. Box 171376 Memphis, TN 38187

Terra C-40 Terra 207 Terra 292 Terra 302 Terra 366

1995 Mississippi Cotton Variety Trials

Varieties submitted for testing were divided into two groups based on maturity classifications determined by the companies submitting each variety. The Early Maturing Cotton Variety Test was comprised of 26 varieties in the Delta and 24 varieties in the Hills. The Mid-season Cotton Variety Test was comprised of 17 varieties in the Delta and 15 varieties in the Hills. Stoneville LA 887 and Deltapine 50 were used as check varieties in tests of both maturity groups.

The Early Maturing and Mid-season Variety Tests were conducted at five locations in the Delta: Stoneville, Tunica, Sumner (planted on 32-inch rows), Elizabeth, and Tribbett. Tribbett was a new site located on a Dundee silty clay, a soil very similar to a Forestdale silty clay. The Early Maturing and Mid-season Tests were conducted at six Hill locations: Mississippi State, Brooksville (planted on 30-inch rows), Raymond, Holly Springs (grown under no-till conditions), Nesbit, and Verona.

With the exceptions noted above, all tests were planted solid in 38- or 40-inch rows. Each variety was replicated six times at each location. Yield determinations were based on the weight of seed cotton mechanically harvested from two-row plots that ranged from 40 to 50 feet in length. Determination of lint fraction, boll size, seed index (weight in grams of 100 fuzzy seed), and fiber properties were made from hand-picked 100-boll samples or from machine-harvested grab samples from three replications at each location. Samples were ginned on a 10-saw laboratory gin. HVI fiber property determinations were made by the Cotton Fiber Laboratory of Louisiana State University, Baton Rouge, LA.

In addition, large, machine-harvested samples were obtained from the tests at Stoneville and Tribbett for processing on the USDA Ginning Laboratory's micro-gin (closely simulating commercial ginning). These samples were used to determine commercial gin turnout, and fiber samples were evaluated at the Agriculture Marketing Service office in Dumas, AR, for classer's grade and HVI fiber properties. From this information, premiums and gross returns per acre (based on UDSA loan schedule which may not accurately reflect differences in final market values) were calculated for each variety.

In addition to yield and fiber quality, relative maturity differences among varieties is an important factor in selecting which varieties to plant. Maturing a crop early is important in avoiding late-season insect pressure and insuring that the crop can be harvested before the onset of adverse weather. Planting two or more varieties that differ in time to maturity may be valuable in spreading out equipment requirements. Various measures of crop development were taken at several locations to provide information on relative differences in maturity. At Stoneville and Tunica, physiological development was monitored from early July until late August by periodically determining nodes above white bloom (NAWB, the number of nodes from the uppermost first position white bloom to the terminal) in ten plants per plot in three replications. A value of 5 or less for NAWB (NAWB5) is considered an indication of imminent cutout and can be used in deciding when to terminate insect control efforts. The number of days from planting to NAWB5 was calculated for each variety.

The number of nodes above the uppermost first position cracked boll (NACB) can be used as an indicator of harvest readiness. A boll is considered mature if it is four or fewer nodes above a boll which has already cracked. When the crop averages four or fewer NACB, it is usually safe to defoliate. NACB counts were made on ten plants per plot in three replications of tests at Stoneville, Tunica, Mississippi State, Brooksville, and Raymond approximately 130 days after planting.

In addition to the main variety tests discussed above, a separate, combined test comprised of 24 varieties was conducted on 30-inch rows at Stoneville, at Hattiesburg, and in a test with and without a nematicide on a reniform nematode-infested soil near Egremont, MS. At Egremont, five replications were used, the test was planted in a 2 x 1 skip row pattern, and only seed cotton yield was determined.

In all tests, seed of each variety was supplied by the company that submitted the variety for testing. Recommended management practices were followed in each test. The on-farm cooperators decided planting dates, fertilizer rates, amount of supplemental irrigation, defoliation date, insect and weed control strategies, and harvest date. These tests do not encompass all growing and environmental conditions in the state, but they provide a guide to producers in selecting among varieties best suited for their area or growing conditions.

At the bottom of each table are summary statistics that are very important in interpreting the test results. Despite efforts to provide a uniform test environment, all experiments are subject to a certain degree of error due to variation between plots arising from differences in soil type, fertility, insect damage, weed pressure, etc. Therefore, yield potential (and performance with respect to other characteristics) cannot be measured with complete accuracy. By conducting replicated trials we can account for or remove some, but not all, of the effect of non-uniform conditions among plots. As a result, the mean performance of some varieties may be numerically different due to natural variation in the data, but not statistically different when variability in the test is taken into account. The least significant difference (LSD) is a statistic that estimates the smallest difference between two varieties that should be considered something other than natural variation. For example, if the LSD for lint yield in a given trial is 80 lbs/A, varieties that differ by less than 80 lbs/A should be considered equal in yield. In key tables and for key traits, values that are not significantly different from the variety with the highest value (or in the case of NAWB5 and NACB, the lowest value) in the trial are shown in bold type. For example, in Table 1 the highest yielding variety at Stoneville, MS, in 1995 was Stoneville 474 with 1195 lbs/A. The LSD for this data set was 64 lbs/A. All varieties yielding within 64 lbs/A of 1195 (i.e. 1131 or greater) are considered equal in yield to Stoneville 474 and are shown in bold type. The LSD values given in this bulletin are at the 5% probability level which means that 5% or less of the time, the analysis indicates that two varieties are equal when in fact they may be different. The letters "NS" in place of an LSD value indicate that there was no statistical difference among varieties in that trial for the trait in question.

The coefficient of variation (CV) is a measure of relative precision of a given trial and is generally considered to be an estimate of the amount of unexplained variation in that trial. In general, the higher the CV, the less precise is a given trial. The R-squared value is another measure of relative precision. The higher the R-squared value, the more precise a trial is.

Data Tables

Return to MAFES Variety Trials | Cotton | Mississippi Agricultural and Forestry Experiment Station