c07-06-0362 Campbell.indd

Historically, reselection, pedigree, and massselection breeding methods have been used to develop open-pollinated cultivars of upland cotton (Gossypium hirsutum L.). As a result, modern cotton cultivars should have accumulated additive genetic effects with time, while also possessing fewer nonadditive gene effects than obsolete cultivars. A topcross test was conducted to compare the heterotic effects of obsolete and modern cultivars for yield, yield components, and fi ber quality. Signifi cant differences were detected between heterosis values for the modern and obsolete cultivar groups for seed cotton yield, lint yield, lint percentage, and boll weight. No signifi cant heterotic effects were detected for fi ber quality. The obsolete group of cultivars showed average lint yield heterosis values of 34% compared with 23% for the modern cultivars. Both cultivar groups displayed signifi cant, but similar heterosis values for the number of bolls per square meter (17 and 15%, respectively). The major yield component associated with lint yield heterosis for both groups was bolls per square meter, although boll weight heterosis also contributed to lint yield heterosis for the obsolete cultivars. Although modern cultivars produced considerable heterotic effects for yield, this study demonstrates that obsolete cultivars may provide an additional source of nonadditive genetic effects that can be exploited in a hybrid production system. B.T. Campbell, USDA-ARS, Coastal Plains Soil, Water, and Plant Research Center, 2611 W. Lucas St., Florence, SC 29501; D.T. Bowman, Dep. of Crop Science, North Carolina State Univ., Raleigh, NC 28796; and D.B. Weaver, Dep. of Agronomy and Soils, Auburn Univ., Auburn, AL 36849. Received 26 June 2007. *Corresponding author ([email protected]). Published in Crop Sci. 48:593–600 (2008). doi: 10.2135/cropsci2007.06.0362 © Crop Science Society of America 677 S. Segoe Rd., Madison, WI 53711 USA All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permission for printing and for reprinting the material contained herein has been obtained by the publisher. R e p ro d u c e d fr o m C ro p S c ie n c e . P u b lis h e d b y C ro p S c ie n c e S o c ie ty o f A m e ri c a . A ll c o p y ri g h ts re s e rv e d . 594 WWW.CROPS.ORG CROP SCIENCE, VOL. 48, MARCH–APRIL 2008 cotton seed in the United States. These hybrids are targeted for the U.S. pima (G. barbadense L.) production area of California and are interspecifi c hybrids between G. hirsutum and G. barbadense (www.hazerainc.com/hsi, verifi ed 11 Jan. 2008). Outside of the United States, however, China and India have rapidly adopted hybrid cotton production systems and increased yield. According to Dongre and Parkhi (2005), hybrid cotton in India represents approximately 45% of the total production area and accounts for about 55% of India’s cotton production. Dong et al. (2006) reported that hybrid cotton production in China since 2000 covers approximately 20% of the total acreage. In addition, Dong et al. (2004) reported that hybrid Bt cotton in India has increased yield 20% compared with pure-line Bt cotton cultivars. In the United States, hybrid upland cotton production has not successfully emerged to date. Meredith and Brown (1998) reported that the major limiting factor preventing U.S. hybrid cotton production and the use of heterosis is the lack of an effi cient and dependable system for producing F1 or F2 hybrid seed. This limiting factor still exists today. In the 1990s, however, the U.S. cotton industry began marketing F2 seed for commercialization following reports by Meredith and Bridge (1972) and Olvey (1986) indicating the feasibility of F2 hybrids. Unfortunately, the commercialization of F2 hybrid seed in the United States did not prove to be a success, mainly due to the ineff ectiveness of the male gametocide, TD-1123 (Meredith and Brown, 1998). Nonetheless, the prospect of commercializing heterosis for U.S. cotton production continues to be appealing today because of success in China and India and well-known heterotic eff ects for yield. In the 1990s, the breeding approach used to capture yield heterosis was based on choosing the highest yielding and most productive parents to develop hybrid seed. Davis (1978) stated that the highest yielding hybrids usually result from crosses involving the highest yielding cultivars. Selecting parental lines to produce hybrids in this way would indicate the ability to accumulate additive genetic eff ects. As such, studies have demonstrated a close relationship between parental performance and that of their hybrids (Miller and Lee, 1964; Wu et al., 2004). Meredith and Brown (1998) reported, however, that unexplained variability, due primarily to nonadditive genetic eff ects, would hinder choosing hybrid parents based on parental performance alone and suggested selecting parents based on their known combining ability. Combining ability has been investigated in cotton for specifi c crosses in numerous studies during the last 50 yr. These studies demonstrate the importance of selecting parental lines based on their individual combining ability to maximize the probability of successful genetic improvement. Overall, these studies were performed to determine the nature of gene action for specifi c traits within specifi c populations. The nature of gene action is important to consider, because heterosis is due to the accumulation of nonadditive genetic eff ects that can result from dominance, partial dominance, overdominance, or epistasis. Young and Murray (1966), Marani (1968), AlRawi and Kohel (1969), Meredith and Bridge (1972), and Tang et al. (1993a) reported that dominance was the predominant form of gene action responsible for yield and yield component heterosis in F1 hybrids. Several studies also reported the presence of epistatic gene eff ects on heterosis expression (Lee et al., 1967; Al-Rawi and Kohel, 1969; Meredith and Bridge, 1972; Meredith, 1990), while other studies comparing F1, F2, and F3 hybrids reported the predominance of nonadditive eff ects in F1 hybrids followed by increased additive genetic eff ects in F2 and F3 hybrids (Meredith and Bridge, 1973; Tang et al., 1993a). The cumulative results of these studies prompted cotton breeders to concentrate eff orts to accumulate additive genetic eff ects for genetic improvement until the development of a feasible hybrid seed production system (Lee et al., 1967; Meredith and Bridge, 1972). Subsequently, cotton breeders have been selecting cultivars from open-pollinated populations since the early 1900s, primarily using reselection, pedigree, and mass selection (Calhoun et al., 2006). These breeding schemes have accumulated additive gene eff ects and may have inadvertently reduced nonadditive gene eff ects that have contributed to a decline in yield heterosis. If this were true, the breeder wishing to utilize heterosis might attempt to incorporate nonadditive gene eff ects from obsolete lines into more agronomically desirable parents. Several studies have compared the mean performance of modern and obsolete cultivars with the objective of determining the rate of genetic gain for yield with time (Bridge et al., 1971; Bridge and Meredith, 1983; Culp, 1984). These studies demonstrated that modern cultivars produce higher yields primarily by increasing lint percentage and bolls per square meter while reducing boll weight (Bridge et al., 1971; Bridge and Meredith, 1983; Culp, 1984). In addition, a series of studies demonstrated that modern cultivars produce higher yields due to their ability to transition into reproductive growth earlier in the growing season and to partition more dry matter resources into reproductive structures (Wells and Meredith, 1984a,b,c). To our knowledge, there are no reports directly comparing the heterotic eff ects of obsolete and modern cotton cultivars. White and Richmond (1963) studied heterosis among crosses of primitive, foreign, and cultivated cottons and found only two instances of heterosis for yield, both involving an old Cambodian type. The objective of this study was to compare the eff ects of heterosis resulting from crosses involving several obsolete, ancestral cultivars and crosses resulting from several modern cultivars. We hypothesize that more nonadditive gene R e p ro d u c e d fr o m C ro p S c ie n c e . P u b lis h e d b y C ro p S c ie n c e S o c ie ty o f A m e ri c a . A ll c o p y ri g h ts re s e rv e d . CROP SCIENCE, VOL. 48, MARCH–APRIL 2008 WWW.CROPS.ORG 595 tilization, weed control, insect control, and defoliation measures were managed following established local practices. Plant growth regulators were not used. Agronomic data were collected on plant height, boll weight, lint yield, lint percentage, bolls per square meter, date of fi rst fl ower, and earliness as measured by the percentage of bolls open 2 wk before harvest. Fiber quality data for fi ber length, strength, micronaire, and uniformity index were measured using high-volume instrumentation analyses. A 25-boll sample was collected before harvest to determine lint percentage, boll weight, and fi ber properties. In 2006, a set of three fi eld trials was conducted in Auburn, AL, Florence, SC, and Rocky Mount, NC. Each of the three trials was designed and conducted in much the same manner as the 1995 study and incorporated a randomized complete block design that included four replicates of the 10 F1 hybrids and 11 parental lines. Similar to 1995, data collected included the agronomic traits seed yield, lint percentage, lint yield, boll weight, and bolls per square meter. No data were collected on plant height and earliness. High-volume instrumentation fi ber quality data were collected for the fi ber properties length, strength, uniformity index, elongation, micronaire, and short fi ber content. Data Analysis In the 1995 trial, heterotic eff ects were calculated for each trait and entry by subtracting midparent values from F1 values in each replicate. The midparent value was calculated as the mean of Georgia King and each F1’s respective parent. For each trait, heterosis percentage was determined by dividing heterotic eff ects by the midparent values and multiplying by 100. Heterosis percentage values were subjected to analysis of variance by PROC GLM to test if diff erences existed between modern and obsolete cultivars for each agronomic and fi ber quality trait (SAS Institute, 2002). In the 2006 trials, entry means for each trait were calculated in each of the three environments. Mean data were used to calculate the midparent, heterotic eff ects, and heterosis percentage values for each of the fi ve modern and fi ve obsolete cultivars. Mean trait data and heterosis percentage values were subjected to analysis of variance by PROC GLM using the following model: = μ+ + + + ε ( ) ijk i j k j ijk Y l g c where Yijk is the mean value of the ith location of the jth group of the kth entry, μ is the overall mean, li is the eff ect of the ith location, g j is the eff ect of the jth group, ck( j) is the eff ect of the kth entry in the jth group and ε ijk is the random error. The ε ijk values are assumed to be independently distributed with constant variance. The LSD for each trait was calculated among all parents (including the tester Georgia King), the 10 F1 hybrid combinations, the 10 midparent values, and the 10 heterosis values. The LSD allowed individual entry comparisons for each trait. Coeffi cients of parentage between and among the tester (Georgia King) and each cultivar (obsolete and modern) included in the 1995 and 2006 trials were calculated as published by Bowman et al. (1997). Coeffi cients for cultivars included in the 2006 trials are provided in Table 1. For each set eff ects exist in crosses involving obsolete cotton cultivars than modern cotton cultivars. MATERIALS AND METHODS Plant Materials and Cultivar Descriptions In a preliminary study in 1995, Georgia King, a modern cultivar, was topcrossed with fi ve modern cultivars (Carolina ES300, Deltapine 50, Deltapine 90, DES 119, and S-35) and 10 obsolete cultivars (Acala 5675, Delfos, Half and Half, Hopi Moencopi, Kekchi, Lightning Express, Lone Star, Rowden, Trice, and Wannamaker’s Cleveland) to produce F1 seed. Georgia King was released in 1990 and derived from a cross between ‘Tifcot 56’ and ‘McNair 235’. Carolina ES-300 is a blend of Coker cultivars, e.g., Coker 310, Coker 312, etc., and was released in 1992. Deltapine 50 was released in 1984 and Deltapine 90 in 1981; DES 119 was made available in 1985 and S-35 in 1989. The choice of Georgia King as the topcross tester was arbitrary. Acala 5675, a reselection of ‘Acala 5’, was released in 1941; Acala 5 dates back to 1917 (Calhoun et al., 2006). A series of Delfos cultivars, which were reselections from ‘Foster’, were off ered beginning in 1920. Half and Half was released in 1905 as a reselection of ‘Cook Improved’. Hopi Moencopi was an ancient cultivar of the Hopi Indians of Arizona and was collected in the early 1930s. Kekchi was introduced from Mexico in 1904. Lightning Express was a reselection of ‘Express 350’ released in 1923. Lone Star was a reselection of ‘Jackson Round Boll’ released in 1905. Trice was also released in 1905 as a reselection of ‘Tennessee Green Seed’. Rowden was a reselection of ‘Bohemian’ released in 1900. Wannamaker’s Cleveland was released in 1915 as a reselection of ‘Cleveland’. A second set of topcrosses was made in 2005 with the same topcross tester Georgia King. All F1 hybrid seed was produced by hand emasculation and hybridization during the summer of 2005 in North Carolina, South Carolina, and Alabama. In this set of topcrosses, fi ve modern cultivars were used and included Deltapine 51, Deltapine 90, Delta Pearl, FiberMax 966, and SureGrow 747. In addition, fi ve obsolete cultivars were used and included Half and Half, Hopi Moencopi, Lone Star, Rowden, and Young’s Acala. Deltapine 51 is a reselection of Deltapine 50, which was used in the preliminary study. Delta Pearl resulted from a cross of ‘Deltapine 5816’ and ‘Sicala 34’. SureGrow 747 has 87.5% DES 119 in its pedigree, which was used in the preliminary study. FiberMax 966 was developed in Australia and made commercially available in the United States in 2000. Young’s Acala is an accession from Mexico introduced into upland cotton breeding programs in the early 1900s as a potential source of boll weevil (Anthonomus grandis Boheman) resistance. Field Trials In a preliminary study, a fi eld trial was conducted in 1995 at the Central Crops Research Station near Clayton, NC. The trial included two replicates of 31 entries each evaluated in tworow plots, 11.1 m long. Row width was approximately 1 m. Field plots were arranged with each F1 hybrid paired between its respective parents and these sets, randomized within blocks. The trial was planted 5 May and harvested 26 October. FerR e p ro d u c e d fr o m C ro p S c ie n c e . P u b lis h e d b y C ro p S c ie n c e S o c ie ty o f A m e ri c a . A ll c o p y ri g h ts re s e rv e d . 596 WWW.CROPS.ORG CROP SCIENCE, VOL. 48, MARCH–APRIL 2008 of cultivars included in the 1995 and 2006 trials, Pearson correlations were calculated to examine the relationship between genetic distance from the tester as estimated by coeffi cients of parentage and heterosis for each trait, with signifi cant diff erences within the modern or obsolete cultivar groups. RESULTS AND DISCUSSION