Upland Regional Variety Trials - 2000

G. Hart and H. Moser

Abstract

Each year the University of Arizona conducts upland cotton variety tests to evaluate the performance of a diverse set of experimental and commercial varieties in Arizona. One such program is the Regional Variety Test (RVT). In 2000, we evaluated a total of 61 varieties at one or more locations in Arizona. These varieties were submitted to us for testing by 13 private seed companies and three public breeding programs. This report presents the results of the trials conducted at Maricopa, Marana, Safford and Yuma.

PDF version

Introduction

Variety selection is the first and perhaps most critical decision cotton producers make each year. A productive variety provides a solid foundation for building a profitable cultural program. Conversely, even the best growers find it difficult to coax profitable yields from a poor variety that is not adapted to the region.

Many sources of information are available to growers to help them choose a good variety. The best source of information, of course, is personal experience with a particular variety on the farm, but given the large number of varieties to choose from, it is impossible to try them all. Many growers, therefore, consider performance data from variety trials that are conducted by seed companies and universities within the region of interest. University variety trials provide growers with unbiased head-to-head comparisons of many varieties from different seed companies. Data generated through these testing programs also are useful to seed companies and other segments of the cotton industry.

The University of Arizona conducted three different upland cotton variety testing programs in 2000- the Advanced Strains Test, the Regional Variety Test, and the State-wide Upland Cotton Variety Test. The Regional Variety Test (RVT) is designed to evaluate a diverse group of commercial varieties in trials that are conducted on research stations throughout the cotton producing regions of Arizona. The RVT is part of a belt-wide program, the National Cotton Variety Testing Program, which is conducted through a cooperative effort among the USDA and various university programs.

The purpose of this report is to present our results from the 2000 Regional Variety Test.

Materials and Methods

We evaluated a total of 61 varieties in one or more trials in 2000. Thirteen private seed companies and three public breeding programs entered varieties into the trials. Four of the varieties; Maxxa, NUCOTN 33B, Atlas, and SG747 are national standards.

We planted trials at four University of Arizona research stations in 2000: Yuma Valley Ag Center (YAC), Maricopa Agricultural Center (MAC), Marana Ag Center (MAR) and Safford Ag Center (SAC). We planted two tests at MAC; an early season test (harvested 15 September) and a full season test (harvested 29 September). The early season test was terminated when SG747 reached maturity, while the full season test was terminated when NuCOTN 33B reached maturity.

The basic experimental procedures were similar for all trials. Plots at MAC, MAR and YAC were four rows wide spaced 40 inches apart and plots at SAC were two rows wide spaced 36 inches apart. The plots varied from 37’ to 46’ feet long. Plots were arranged in a randomized complete block design with five replications at MAC and MAR and four replications at YAC and SAC. We used a high seedling rate of approximately 30 lbs/A and then thinned all plots to a final stand of three plants per foot of row. We planted the YAC test on 10 March, the two MAC tests on 13 April, the MAR test on 20 April and the SAC test on 21 April. The target population of approximately 40,000 plants per acre was reached for most of the plots, however a few of the plots with unacceptable stands were discarded from the data set.

All tests were grown using standard herbicide and fertility regimes. We used IPM practices for conventional varieties to control pink-bollworm, lygus and whitefly in all five trials.

The plots at MAR were visually evaluated for foliar symptons of Verticillium wilt just prior to defoliation. Each plot was given a score ranging from 0 for no visible symptoms to 9 for extensive foliar symptoms. The plots at MAC and YAC were visually evaluated for extent of lodging with 1= no lodging and 5=extensive lodging.

A 50 boll sample from each plot in three replications at each site was picked just prior to machine harvest. Boll size, lint percent and seed percent were determined from this sample and the lint was sent to ITC (International Textile Center) for HVI analysis. The middle two rows of each 4-row plot were mechanically harvested for seed cotton weight. The lint yield was calculated from the lint percent obtained by the 50 boll samples. Plant height was measured from three replications of each test immediately after harvest. The plots were harvested on 7 September at YAC, 15 September for the early season test at MAC, 29 September for the late season test at MAC, 5 October at SAC and 20 November at MAR. The late date for MAR harvest was due to unusual heavy rain in October.

We conducted an analysis of variance on the data for each test and calculated protected Least Significant Difference (LSD) values for all applicable traits.

Results and Discussion

In the early season test at MAC, the average lint yield of the 28 entries was 1523 lbs/A. The range was from 2002 (SG747) to 1191 lbs/A (PG094) with SG747 yielding significantly more than the other entries. The value per acre($/A) and the premium(cents/lb) are listed in Table 1. A negative number in the premium column denotes a discount on the price paid. Agronomic characteristics for the early season at MAC is listed in Table 2. Note that both the full season and the early season tests experienced some lodging shortly after peak bloom.

The mean for the full season test at MAC was slightly higher than the early season test at 1639 lbs of lint per acre. The range was 2030 (SG747) to 1207 lbs/A (Atlas). As in the early season test, SG747 yielded significantly higher than the other 23 varieties (Table 3). SG747 had the greatest value per acre even though it was heavily discounted because of fiber quality.

In the test at MAR, the average lint yield was lower at 1144 lbs/A due to the cotton stringing out after heavy rains in October. PM1560BR, AP9257, SG747, SG501BR and STX9903RR yielded significantly higher than the other 31

varieties (Table 5). AP9257 and PM1560BR produced the highest value per acre. The correlation between yield and Verticillium wilt susceptibility was not as apparent as it was last year (Table 6).

The average lint yield at SAC of the 43 entries was 1030 lbs/A. The range was from 1329(PM1560BR) to Atlas (608). PM1560BR, DP655, DP458BR, Delta Pearl and NUCOTN33B produced the highest yields (Table 7). Agronomic characteristics for the SAC trial are listed in Table 8.

The average lint yield at YAC of the 28 entries was very good at 1941 lbs/A (Table 9). The range was from 2466 (SG747) to 1394 (PG094). SG747, SG215BR, and SG501BR were the highest yielding varieties in this test. The variety with the highest value, however, was SG125BR because of the higher premium paid on the fiber qualities relative to SG747 and SG215BR. Agronomic characteristics are listed in Table 10. Lodging of some varieties in this test was extensive due to strong winds that struck shortly after peak bloom.

2000 was a good year for cotton production in Arizona. The only serious weather problem was a very wet October in Marana, which delayed harvest and probably decreased the yields in this test. Heat stress was present in the low desert elevations, but was not unusually high for Arizona. Insect pressure in all five tests were light.

Acknowledgements

We wish to thank the seed companies for their cooperation and support of the variety testing program at the University of Arizona. The land and resources provided by the University of Arizona Agricultural centers at Yuma, Maricopa, Marana and Safford are also greatly appreciated.