RESEARCH

Germplasm Potential for Continuing Improvement of Fiber Quality in Upland Cotton: Combining Ability for Lint Yield and Fiber Quality
Linghe Zeng,* William R. Meredith Jr., and Deborah L. Boykin
USDA-ARS, Crop Genetics Research Unit, Delta Research Center, Stoneville, MS 38776. Mention of trade names or commercial products in this article is solely for the purpose of providing specific information and does not imply recommendation by the USDA. Received 14 July 2010. *Corresponding author (linghe.zeng@ars.usda.gov). Abbreviations: GCA, general combining ability; HPH, high parent heterosis; JC, John Cotton; SCA, specific combining ability; SP, Species Polycross.

ABSTRACT
Exotic germplasm, that is, germplasm without commercial applicability, can broaden the genetic base in upland cotton (Gossypium hirsutum L.). The objectives of this study were to determine combining ability and identify parents from exotic germplasm for breeding. Twelve regionally adapted exotic germplasm lines were crossed with one cultivar and three elite germplasm lines in a North Carolina Design II scheme. The 48 F2 hybrids were evaluated at two locations with 4 and 3 replicates each in 2008 and 2009, respectively. General combining ability (GCA) effects were significant (p 0.001) for lint yield and all fiber properties. General combining ability effects were more important than specific combining ability (SCA) effects for most traits. Significant (p 0.05) favorable GCA effects were detected in all exotic germplasm parents for different fiber properties. Positive GCA effects for lint yield and favorable GCA effects for at least one fiber property were detected in some exotic parents: SP156 and SP224 for lint yield and strength; SP192, SP205, and JC65 for lint yield, micronaire, and elongation; and SP192 and SP224 for lint yield and short fiber content and fineness, respectively. Significant high parent heterosis (HPH) for lint yield was detected in F2 hybrids of FM832 SP205 (14%) and MD15 SP205 (26%). These results provide evidence for the potential of these germplasm lines in breeding for continuing improvement of lint yield and fiber quality.

xploration of exotic germplasm resources and introgression of exotic genes into commercial cultivars in upland cotton (Gossypium hirsutum L.) provide an appealing approach to broaden the genetic base and genetically improve cultivars for lint yield and fiber quality. Previous efforts from cotton breeders in the United States have incorporated day-neutral genes into tropical germplasm landraces (McCarty et al., 1979) that have facilitated the utilization of these primitive accessions. However, the problem associated with recovery of exotic alleles during conversion of day-neutral genes, because of linkage drag from the day-neutral donor parent (Liu et al., 2000), may limit a wide use of these germplasm. Recently, some regionally adapted exotic germplasm populations derived from crosses among tetraploid species of Gossypium have been evaluated (Zeng et al., 2007; Zeng and Meredith, 2009a). These germplasm populations were initiated by the U.S. cotton breeders in 1960s and 1970s and underwent long term introgression and inbreeding since their initiation. The most important feature of these germplasm resources is their stabilized genomes and, therefore, the problems commonly associated with interspecific hybrids such as linkage drag (Young and Tanksley, 1989) and segregation distortion (Jiang et al., 2000) can be at least partially overcome in these germplasm resources. Dramatic genetic variations in lint yield

Published in Crop Sci. 51:6068 (2011). doi: 10.2135/cropsci2010.07.0413 Published online 15 Nov. 2010. Crop Science Society of America | 5585 Guilford 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.

60

WWW.CROPS.ORG

CROP SCIENCE, VOL. 51, JANUARY FEBRUARY 2011

and fiber properties have been identified in these germplasm populations. Further utilization of these germplasm resources in breeding requires information about combining ability for lint yield and fiber quality when the exotic germplasm lines are used as parents in breeding. Combining ability for lint yield and fiber properties has been extensively analyzed and numerous studies have been reported since the middle of the last century. Miller and Marani (1963) analyzed 28 F1 and F2 hybrids in a half diallel design and identified significant general combining ability (GCA) and specific combining ability (SCA) for lint yield and fiber properties although GCA effects were more important than those of SCA. The interactions of GCA and location were not significant in their study. In the same study, the rankings of the lines based on GCA were similar to those based on parental performance, which indicated the predictability of parental performance by their GCA. Lee et al. (1967) and El-Adl and Miller (1971) detected significant GCA for lint yield and fiber properties with no significant GCA environment in F1 hybrids. In contrast, Meredith and Bridge (1973) detected significant SCA for lint yield in analysis among 10 F2 hybrids. More recently, Tang et al. (1993a, b) reported significant GCA and SCA for lint yield and fiber properties and significant interactions of GCA and environment for most traits analyzed in 64 F2 hybrids. Meredith and Brown (1998) analyzed 120 F2 hybrids in a half diallel design and reported that SCA and SCA location accounted for 79% of total variation for lint yield. In summary, GCA effects were generally more important than SCA effects in most studies except the report by Meredith and Brown (1998). The presence of large additive effects suggests that progress of selection for lint yield and fiber properties can be expected from these parents and early generations of segregating populations with these parents. Inconsistent results related to interactions with environment were caused by the diverse genetic backgrounds of those parental lines and different environments. Interrelationships between lint yield and fiber quality have been extensively analyzed (Miller and Rawlings, 1967; Meredith and Bridge, 1972; Culp and Harrell, 1975; Meredith, 1984; Smith and Coyle, 1997; Percy et al., 2006). As reviewed in a recent study by Zeng and Meredith (2009a), lint yield was correlated negatively with fiber strength while positively with micronaire and elongation. Correlation between lint yield and fiber length was either negative (r = 0.47) (Meredith and Bridge, 1972) or nonsignificant (Miller and Rawlings, 1967; Percy et al., 2006). Genotypic correlation coefficients between lint yield and short fiber content and between lint yield and fineness were 0.08 and 0.57, respectively, based on analysis in 200 exotic germplasm lines derived from interspecific crosses between G. hirsutum and G. barbadense L. (Zeng and Meredith, 2009a). Success in breaking the negative association between lint yield and fiber quality has been limited in the past. Green and Culp
CROP SCIENCE, VOL. 51, JANUARY FEBRUARY 2011

(1990) detected positive GCA effects for both lint yield and yarn strength in one cultivar, SC-1 (Culp and Harrell, 1979). In another study of 15 F1 hybrids by a half diallel design (Coyle and Smith, 1997), GCA effects for fiber properties among genotypes were unfavorably associated with GCA for within-boll-yield components. Identification of favorable GCA effects for both lint yield and fiber properties in parents will help selection of parents in breeding for simultaneous improvement of lint yield and fiber quality. Knowledge on combining ability for lint yield and fiber quality between the exotic and elite germplasm lines will help enhance the utilization of these germplasm lines in cotton breeding. A breeding scheme was designed to cross regionally adapted exotic germplasm lines with cultivar and elite germplasm lines with good fiber quality to introgress exotic genes into upland cotton for continuing genetic improvement of fiber quality. The objectives of this study were to determine combining ability effects in F2 hybrids, identify good combiners in exotic germplasm for lint yield and fiber quality, and detect useful heterosis for lint yield and fiber quality in F2 hybrids. Results were expected to identify parents with favorable GCA effects for lint yield and fiber properties, reveal gene action for combining ability effects in the exotic germplasm, and provide implication in breeding and therefore, facilitate incorporation of these exotic germplasm lines into upland cotton. It was also expected that information about combining ability effects in F2 hybrids would reveal possibility for simultaneous improvement of lint yield and fiber quality during introgression of these exotic germplasm in cotton breeding programs.

MATERIALS AND METHODS

One cultivar, Fibermax 832 (FM832; Constable et al., 2001) and three elite germplasm lines, TAM98D-99ne (PI636491; Thaxton et al., 2005), MD52ne (PI 634930; Meredith, 2005), and MD15 (PI 642769; Meredith, 2006) were crossed with 12 exotic germplasm lines in a North Carolina Design II mating design. These cultivar and elite germplasm lines were selected as parents for crossing because of their high fiber quality as reported in their original regions. Male parents were twelve exotic germplasm lines: SP156, SP164, SP192, SP205, SP224, SP235, JC14, JC32, JC60, JC65, JC84, and JC186. Among them, lines from the Species Polycross (SP) designation were derived from multiple crosses among five tetraploid Gossypium species (Zeng et al., 2007) and those from John Cotton (JC) designation were derived from multiple crosses between G. hirsutum and G. barbadense (Zeng and Meredith, 2009a). Six of these lines, SP156 (PI6545087), SP205 (PI654090), JC14 (PI658308), JC32 (PI658309), JC60 (PI658310), and JC65 (PI658311), have been released and deposited in the National Plant Germplasm System (Zeng and Meredith, 2009b; Zeng et al., 2010). These twelve exotic lines were selected as parents for crossing because of their high fiber quality and moderate lint yield. Crosses were made in summer 2007 at the Delta Research Center, Stoneville, MS. The F1 hybrids were grown and selfpollinated at a winter nursery at Colima, Mexico. The 48 F2 hybrids and their 16 parents were evaluated at two sites at the Delta 61

WWW.CROPS.ORG

Research Center, Stoneville, MS during 2008 and 2009 with four replicates at each site in 2008 and three replicates at each site in 2009. The experimental design was a randomized complete block. Plants were grown in single-row plots 12.2 m 1.0 m. Field site 1 and field site 2 were located about 1000 m apart with soil types of Beulah fine sandy loam (a coarse-loamy, mixed, active, thermic Typic Dystrochrepts) for field site 1 and Bosket fine sandy loam (a fine-loamy, mixed, active, thermic Mollic Hapludalf) for field site 2. Seeds were planted on 21 Apr. 2008 at site 1, 30 Apr. 2008 at site 2, 22 Apr. 2009 at site 1, and 23 Apr. 2009 at site 2. Standard conventional field practices were applied in both years and both sites. The factor of environment was considered as a replacement for factors of year and location for purpose of statistical analysis: The first environment, Environment 1, was assigned to site 1 in 2008; the second environment, Environment 2, was assigned to site 2 in 2008; the third environment, Environment 3, was assigned to site 1 in 2009; the fourth environment, Environment 4, was assigned to site 2 in 2009. All 64 entries including 48 F2 hybrids and 16 parents were assigned randomly to each replicate in all environments. At harvest, 50 open bolls from each plot were hand harvested randomly. The remaining bolls from each plot were harvested by a mechanical picker for yield measurements. Boll samples from individual plots were ginned separately using a laboratory saw gin. Total seed cotton weight of each plot was sum of seed cotton weight of the sampled bolls and the remaining bolls in that plot. Each boll sample was used to determine lint percentage measured as the percentage of lint weight from the boll sample divided by the seed cotton weight of the boll sample. Lint yield of each plot was calculated from seed cotton weight per plot and lint percentage of that plot. Twenty grams of lint for each sample were submitted to StarLab (Knoxville, TN) to determine fiber quality. Fiber strength (kN m1 kg1) was measured by a stelometer as the force required for breaking a bundle of fibers. Elongation was the percentage of elongation at the point of break in strength determination. Fiber span lengths were measured as mean length of the longest 50% or 2.5% of the fibers scanned. Micronaire was measured in micronaire units using the Fibronaire instrument (Motion Control, Inc., Dallas, TX). Fibers were also analyzed for mean short fiber content and fineness using the Advanced Fiber Information System (Uster AFIS, 1977). For these measurements, two samples of 0.5 g lint each were taken from each plot. Short fiber content was measured as the percentage by weight of the fibers that were less than 12.7 mm. Fineness was measured as the weight per unit of length (mg km1) where smaller values indicate a higher degree of fineness in fibers. Analysis of variances was performed using the GLM procedure of the Statistical Analysis System (SAS Institute, 2004). A mixed model was used with genotype as a fi xed effect and environment and replicate within environments as random effects. Sum of squares for genotype was partitioned into variation due to parent, parents vs. hybrids, and hybrids. Sum of squares of hybrids was further partitioned into variation due to females, males, and female male. Main effects of the females and males are equivalent to GCA effects of the female and male parents, respectively, and the female male interaction is equivalent to SCA (Hallauer and Miranda, 1981). Combining ability effects of parents were calculated using only the F2 populations as described by Simmonds and Smartt (1999). Standard errors for GCA effects of females and males 62

were calculated followed the method described by Cox and Frey (1984). The significance for each GCA effect was tested in two-tailed t test (Cox and Frey, 1984). Mean separation among parents was performed using the LSD test with error mean squares from parents (df = 150). Means of the 48 F2 populations were compared with means of the 16 parents in t test comparisons. Mean squares from error in the F2 populations (df = 470) and parents (df = 150) were used as variances in the t tests. Means of the 24 F2 hybrids derived from SP germplasm were compared with means of the 24 F2 hybrids derived from JC germplasm in t test comparisons. Similarly, mean squares from error (df = 230) were used as variances in the t tests.

RESULTS AND DISCUSSION

Genotype effects were highly significant (p 0.001) for all the traits and interactions of genotype and environment were significant (p 0.05) for most traits except fiber elongation and 50% span length (Table 1). Total variation among the 48 F2 hybrids is partitioned into GCA and SCA effects (Table 1). Effects of GCA estimated from female parents and male parents were highly significant for all traits. Specific combining ability was highly significant for micronaire, elongation, span lengths, and fineness but not significant for the remaining traits. For those traits with significant SCA effects, GCA effects were predominant based on mean squares. For example, mean squares of fineness for GCA from female and male parents were 123 and 19 times larger than that for SCA, respectively. Sprague and Tatum (1942) defined GCA as the average performance of a line in hybrid combinations and SCA as the deviation of a certain cross from the expected, based on average performance of the parental lines involved. They further interpreted GCA and SCA as largely additive effects and nonadditive effects of dominance and epistasis, respectively. The significant GCA and SCA in F1 hybrids resulted from differences of additive and nonadditive effects (Beil and Atkins, 1967). Since 50% dominance is expected to be lost from F1 to F2 generation, it would be reasonable to expect less nonadditive effects in F2 hybrids than in F1 hybrids. In studies by Tang et al. (1993a, b), additive effects were found to be more important than nonadditive effects for lint yield and fiber properties in 64 F2 cotton hybrids. In another study by Meredith and Brown (1998), SCA and SCA location accounted for the majority of the variation for lint yield of 120 F2 cotton hybrids. The results in these studies confirmed the arguments that estimation of GCA and SCA depends on the particular set of parents in hybrid test (Hallauer and Miranda, 1981). The results in this study that additive effects were more important than nonadditive effects for lint yield and fiber properties using F2 hybrids are consistent with those reported by Tang et al. (1993a, b). These results implied that selection in early generations such as the F2 involving SP and JC germplasm could be expected to improve lint yield and fiber quality.
CROP SCIENCE, VOL. 51, JANUARY FEBRUARY 2011

WWW.CROPS.ORG

Table 1. Mean squares of cotton F2 hybrids and the parental lines for lint yield and ber properties when grown in four environments at Stoneville, MS.
Source
Environments (E) Replicate Genotype (G) Parents Female vs. male Female (F) Male (M) Parents vs. hybrids Hybrids GCA (female) GCA (male) SCA GE Parent E (F vs. M) E Female E Male E (Parent vs. hybrid) E Hybrid E GCA (female) E GCA (male) E SCA E Error

df
3 10 63 15 1 3 11 1 47 3 11 33 189 45 3 9 33 3 141 9 33 99 630

Lint yield
475*** 4.4* 45*** 62*** 672*** 27*** 17*** 731*** 24*** 89*** 72*** 2.6 4*** 6.2*** 12*** 5.4** 5.8*** 1.6 3.3*** 13*** 5.0*** 1.9 2.2

Micronaire Elongation Strength

5.0*** 0.44*** 1.0*** 1.2*** 1.3*** 2.1*** 0.99*** 0.12 0.98*** 8.1*** 1.5*** 0.15*** 0.07*** 0.09*** 0.37*** 0.04 0.07 0.22** 0.07** 0.13** 0.08** 0.06 0.05 8.2*** 7.5*** 8.0*** 14*** 9.4** 3.5* 17*** 2.5 6.4*** 13*** 16*** 2.6*** 1.1 0.72 0.22 0.94 0.70 0.74 1.2 1.3 1.6* 0.98 1.0 65*** 2.0*** 2.0*** 1.7*** 3.8*** 4.3*** 0.91*** 3.6*** 2.1*** 20*** 2.7*** 0.29 0.30* 0.34* 0.36 0.47** 0.30* 0.08 0.29* 0.67** 0.30 0.25 0.23

50% span 2.5% span length length

46*** 3.3*** 1.6*** 1.7*** 0.00 1.0* 2.3*** 9.0*** 1.5*** 10*** 1.6*** 0.62*** 0.33 0.42* 0.15 0.36 0.46* 0.38 0.30 0.63* 0.27 0.27 0.29 51*** 4.0*** 10*** 17*** 17*** 15*** 18*** 70*** 7.1*** 41*** 14*** 1.8*** 0.67*** 0.86** 0.78 0.96* 0.84** 0.75 0.61** 2.0*** 0.85*** 0.40 0.42

Short ber content

83*** 10*** 1.49*** 2.2*** 0.79 5.1*** 1.7** 1.3 1.3*** 5.7*** 1.8*** 0.67 0.79* 1.1** 2.3* 0.57 1.1* 2.1* 0.68 1.1 0.69 0.65 0.64

Fineness
3813*** 269*** 762*** 964*** 862*** 1939*** 791*** 32 713*** 6749*** 1042*** 55*** 43*** 45** 129** 38 39* 109** 41*** 61** 49** 37** 25

*Signicant at p 0.05. **Signicant at p 0.01. ***Signicant at p 0.001. Values in this column were decreased from real values 10 4. Values in this column were decreased from real values 10 3. GCA, general combining ability. SCA, specic combining ability.

Interactions of Hybrid and Environment

The mean square of genotype environment can be further partitioned into the mean squares due to parent environment components, parent vs. hybrid environment components, and hybrid environment components (Table 1). The hybrid environment interaction sources of variation were significant for lint yield, micronaire, strength, 2.5% span length, and fineness but were not significant for the remaining traits. The variation of genotype (parent, hybrid, and parent vs. hybrid) environment was mainly attributed to interactions of hybrid and environment because the sums of squares for hybrid environment accounted for the majority portion of the total sums of squares (> 62%) for each trait. This is expected since a wide genotypic spectrum exists in 48 F2 hybrids derived from the 16 parents consisting of elite and exotic germplasm lines. However, less variation of hybrid environment than parent environment for lint yield based on mean squares indicates that agronomic performance of F2 hybrids is more stable than that of parents across environments. Mean squares of hybrid environment and parent environment were similar for fiber properties.
CROP SCIENCE, VOL. 51, JANUARY FEBRUARY 2011

When mean square of hybrid environment is further partitioned into the attributes by GCA and SCA, GCA environment interactions were significant for most traits (Table 1). These results indicate the instability of GCA effects in female or male parents across environments for the respective traits. However, GCA environment interactions in male parents were not significant for strength, 50% span length, and short fiber content. Significant (p 0.01) SCA environment was only observed for fi neness.

Performance of F2 Hybrids and Parents

Significant (p 0.001) overall genotypic differences were observed among parents and hybrids for all traits (Table1). Significant (p 0.01) female vs. male for most traits except 50% span length and short fiber content (Table 1) indicates overall difference in the performance between female and male parents. Lint yield and fiber properties of the parents and F2 hybrids are presented as means across the four environments (Table 2). The general high quality observed among female parents, that is, cultivar and elite germplasm lines, is expected since these
63

WWW.CROPS.ORG

Table 2. Means of lint yield and ber properties in the 16 parents and F2 populations grown in four environments at Stoneville, MS.
Parents
FM832 TAM98D-99ne MD52ne MD15 SP156 SP164 SP192 SP205 SP224 SP235 JC14 JC32 JC60 JC65 JC84 JC186 LSD Mean parents Mean F2 SP-derived F2 JC#-derived F2

Lint yield
kg ha1 1359 1299 1350 1066 994 858 1051 732 985 807 747 888 715 878 948 820 106 969* 1177 1231** 1123

Micronaire
4.58 5.43 5.21 4.44 5.23 4.16 5.29 4.81 4.95 4.99 4.61 4.82 4.56 4.94 4.95 4.53 0.17 4.84 4.81 4.88* 4.74

Elongation
% 5.84 6.75 6.53 7.09 6.78 6.63 7.84 6.78 6.03 6.06 6.31 8.59 6.47 9.25 4.97 7.63 0.76 6.85* 6.69 6.77 6.61

Strength
kN m 1 kg1 242 222 263 256 250 252 225 242 261 242 255 240 241 236 244 237 11.2 244 238 240 235

50% span length

mm 15.8 15.2 15.8 16.2 15.7 16.8 14.5 15.9 15.3 16.1 15.8 15.4 15.8 15.3 15.5 15.8 0.40 15.7 15.6 15.6 15.5

2.5% span length

mm 31.4 29.0 30.7 31.6 30.2 32.7 27.2 30.2 28.9 30.3 30.8 29.9 30.3 29.6 30.2 30.7 0.50 30.2 30.5 30.6 30.4

Short ber content

% 4.24 3.27 2.99 2.96 3.00 3.24 3.71 3.22 3.96 3.00 3.69 3.84 3.58 3.42 3.39 3.96 0.62 3.47* 3.38 3.42 3.33

Fineness
mg km1 172 195 190 173 188 173 194 176 174 183 172 172 171 177 184 172 3.6 179 180 182*** 177

*Signicant at p 0.05. **Signicant at p 0.01. ***Signicant at p 0.001. Multiple comparisons were made among the 16 parents. t tests were made between the means of the 16 parents and the means of the 48 F2 hybrids. SP, Species Polycross. t tests were made between the means of the F2 hybrids derived from SP germplasm and the F2 hybrids derived from JC germplasm. # JC, John Cotton.

lines were selected as parents based on their reputation of fiber quality in their original regions. Among them, MD15 exhibited the best fiber quality with a combination of good properties in strength, elongation, 50% span lengths, 2.5% span length, short fiber content, and fi neness but also produced the lowest lint yield. Male parents, that is, exotic germplasm lines, produced moderate lint yield with diversified fiber properties, for example, 4.9 to 9.25% for elongation, 14.5 to 16.8 mm for 50% span length, and 171 to 194 mg km1 for fi neness. SP205 was released for its moderate lint yield, 1033 kg ha1, and good fiber quality previously (Zeng and Meredith, 2009b). In this study, the lint yield for SP205 was lower than that previously reported due to cold and wet weather before mechanical harvest in 2009. The F2 hybrids averaged significantly higher lint yield than parents (1177 vs. 969 kg ha1) with similar means for fiber quality as those in parents for most properties except elongation and short fiber content. When F2 hybrids were separated into two groups, the SP-derived and JC-derived, lint yield for the SP-derived F2 hybrids was higher (p 0.01) than that for the JC-derived F2 hybrids. Fiber properties were similar between these two groups except micronaire.
64

Combining Ability of Parents for Lint Yield and Fiber Properties

Significant and positive GCA effects for lint yield were detected in six parents: FM832, SP156, SP192, SP205, SP224, and JC65 (Table 3). Among these parents, GCA effects of FM832, SP156, and SP205 were consistent across environments while those of SP192, SP224, and SP65 were inconsistent across environments (data not shown). MD15 was the best general combiner for fiber properties in F2 hybrids with significant and positive GCA effects for elongation, strength, and span lengths and significant and negative GCA effects for short fiber content and fineness (Table 3). Since lower values of short fiber content and fineness mean higher quality in these two traits, the significant and negative GCA values indicate favorable gene effects for these two traits. For convenience, the negative GCA values for short fiber content and fineness and the positive GCA values for other properties were called favorable GCA effects in the following discussion. All male parents, that is, exotic germplasm lines, had significant favorable GCA effects for at least one property while JC60 was the best combiner among male parents with significant favorable GCA for most properties except micronaire and elongation.
CROP SCIENCE, VOL. 51, JANUARY FEBRUARY 2011

WWW.CROPS.ORG

Table 3. General combining ability (GCA) for lint yield and ber properties of the four elite genotypes and twelve exotic germplasm lines grown at Stoneville, MS, across four environments.
Parents
FM832 TAM98D-99ne MD52ne MD15 SP156 SP164 SP192 SP205 SP224 SP235 JC14 JC32 JC60 JC65 JC84 JC186

Lint yield
kg ha1 96.3*** 12.0 35.7 72.5** 95.3*** 0.991 59.0** 225*** 84.4*** 142*** 157*** 5.01 147*** 44.1* 41.9* 26.6

Micronaire
0.047* 0.245*** 0.069** 0.267*** 0.170*** 0.243*** 0.189*** 0.061* 0.049 0.236*** 0.165*** 0.017 0.152** 0.117** 0.062* 0.118**

Elongation
% 0.343*** 0.303*** 0.220** 0.259*** 0.767*** 0.186 0.272* 0.538*** 0.243* 0.077 0.869*** 0.468** 0.681*** 0.866*** 0.038 0.116

Strength
kN m 1 kg1 6.02*** 13.6*** 2.47* 17.2*** 3.24* 0.119 3.43* 13.6*** 5.88** 2.24 9.00*** 4.08* 13.5*** 8.02*** 3.77* 3.51*

50% span length

mm 0.063 0.276*** 0.197*** 0.410*** 0.001 0.426*** 0.229*** 0.059 0.174** 0.190** 0.112* 0.174** 0.300*** 0.301*** 0.152* 0.136*

2.5% span length

mm 0.230* 0.634*** 0.189* 0.594*** 0.601 1.43*** 0.674*** 0.164* 0.503*** 0.019 0.021 0.297** 0.283* 0.486*** 0.122 0.144*

Short ber content

% 0.227*** 0.100* 0.067 0.195** 0.091 0.232* 0.302** 0.277** 0.170* 0.007 0.207* 0.066 0.180* 0.152* 0.050 0.068

Fineness
mg km1 1.70* 8.38*** 0.351 7.02*** 5.24*** 2.54** 8.21*** 0.851 1.61* 5.92*** 5.33*** 1.60* 5.36*** 0.078 1.01 2.70***

*Signicant at p 0.05. **Signicant at p 0.01. ***Signicant at p 0.001.

The GCA effects of JC60 were consistent across all environments. A previous report (Zeng and Meredith, 2009a) showed that short fiber content was negatively correlated with fineness (r = 0.41). However, this study revealed favorable GCA effects for both short fiber content and fineness in JC14 and JC60 (Table 3), thus indicating that these two properties can be simultaneously improved if these germplasm lines are used as parents in future breeding. Although GCA effects were different across environments in some parents for some traits, those with a high significance level, that is, p 0.001, for GCA effects in male parents for lint yield and fiber properties, such as SP156 and SP205 for lint yield, JC65 for elongation, JC14 and JC60 for strength, SP164 for 2.5% span length, and JC14, JC60, and JC186 for fineness were consistent across environments. These results imply that general adaptability could be emphasized in breeding involving these lines for the respective properties. Significant and positive SCA effects for lint yield were detected in only one F2 hybrid, FM832 SP235, and this effect was not consistent across environments (data not shown). Generally, SCA effects were small for most fiber properties except micronaire and span length. Significant and favorable SCA effects were only detected in three F2 hybrids for elongation, four for strength, one for short fiber content, and four for fineness (data not shown). This is consistent with the results in the analysis of variances that indicate limited nonadditive gene effects for these traits in the F2 hybrids. In contrast, significant and positive SCA effects were detected in 15 to 25% of the F2 hybrids for micronaire and span lengths which indicates relatively large nonadditive effects in F2 hybrids for these properties.
CROP SCIENCE, VOL. 51, JANUARY FEBRUARY 2011

Potential of Simultaneously Improving both Lint Yield and Fiber Quality

It is accepted generally among cotton breeders (Miller and Rawlings, 1967; Meredith, 1984; Smith and Coyle, 1997; Percy et al., 2006) that lint yield is negatively associated with fiber quality. However, the demand for high quality raw fibers is increasing due to shift from a domestic to an export market and rapidly rising spinning speed in the surviving U.S. textile industry. Introgression of exotic germplasm into upland cotton may broaden its genetic base for lint yield and fiber quality and increase success in reducing the negative association among yield and fiber traits. In this study, SP156 and SP224 had significant and positive GCA effects for both lint yield and strength (Table 3), which was consistent in all environments. Besides, SP192, SP205, and JC65 had positive GCA effects for lint yield and at least another two fiber properties, for example, micronaire and elongation. Positive GCA effect for lint yield and favorable GCA effect for short fiber content were detected in SP192 and positive GCA effect for lint yield and favorable GCA effect for fineness were detected in SP224 (Table 3). JC186 had favorable GCA effects for strength, 50% span length, 2.5% span length, and fineness without compensation in other traits. Although a negative GCA effect for micronaire was detected in JC186, the value of this property in JC186 was within the range of market standards. These GCA effects of JC186 were consistent generally across all environments tested except for 2.5% span length that was only significant in two of the four environments. In summary, none of the parents had significant favorable GCA effects for lint yield and all fiber properties. Therefore,
65

WWW.CROPS.ORG

Table 4. High-parent heterosis (%) in F2 hybrids for lint yield and ber properties when grown in four environments at Stoneville, MS.
Pedigree
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 FM832 SP156 FM832 SP164 FM832 SP192 FM832 SP205 FM832 SP224 FM832 SP235 FM832 JC14 FM832 JC32 FM832 JC60 FM832 JC65 FM832 JC84 FM832 JC186 TX98-99ne SP156 TX98-99ne SP164 TX98-99ne SP192 TX98-99ne SP205 TX98-99ne SP224 TX98-99ne SP235 TX98-99ne JC14 TX98-99ne JC32 TX98-99ne JC60 TX98-99ne JC65 TX98-99ne JC84 TX98-99ne JC186 MD52ne SP156 MD52ne SP164 MD52ne SP192 MD52ne SP205 MD52ne SP224 MD52ne SP235 MD52ne JC14 MD52ne JC32 MD52ne JC60 MD52ne JC65 MD52ne JC84 MD52ne JC186 MD15 SP156 MD15 SP164 MD15 SP192 MD15 SP205 MD15 SP224 MD15 SP235 MD15 JC14 MD15 JC32 MD15 JC60 MD15 JC65 MD15 JC84 MD15 JC186

Lint yield
4.6 8.9* 7.5 14** 0.85 16*** 12* 3.9 19*** 6.0 5.6 5.8 0.30 8.6 2.3 3.2 1.1 18*** 23*** 7.5 21*** 4.9 10* 8.6 7.5 15*** 7.2 2.4 9.7* 24*** 27*** 16*** 26*** 13*** 22*** 22*** 17** 6.3 9.3 26*** 12* 15*** 5.7 6.7 1.8 13* 2.8 0.38

Micronaire Elongation
2.2 2.6 4.4*** 5.3* 1.8 1.6 7.2*** 5.0* 1.5 0.81 5.6 2.4 0.37 8.8*** 0.10 3.5*** 4.5*** 1.9 10*** 6.5*** 6.5*** 1.2 4.9*** 8.1*** 2.3 6.9*** 0.57 6.1*** 6.4*** 3.7* 4.9*** 5.0*** 7.6*** 3.1** 5.0** 3.6** 8.7*** 0.79 11*** 1.0 11*** 0.46 3.5*** 2.7* 0.00 1.6 5.6*** 0.22 23*** 0.01 22*** 2.3 4.2 10 13 19*** 11** 21*** 9.7 16* 7.4 7.0 7.1* 20** 4.2 4.6 9.3 16* 12** 16* 4.6 5.8 12 2.8 13* 6.5 5.7 0.92 6.7 20** 13 25*** 15 9.4 13* 4.4 3.9 3.0 3.9 4.4 12* 13 6.6 12 11 13*

Strength
2.0 2.0 1.5 6.1*** 3.8 2.2 3.8 2.7 4.8*** 1.7 1.1 1.5 5.2 5.6 3.0 3.8 8.3 4.5 3.2 2.5 2.2 4.9* 2.1 2.2 2.8 5.3** 7.9*** 11*** 1.5 3.1 0.76 4.8* 1.3 6.3** 8.0*** 1.3 7.0 3.6 6.4 2.8 4.4** 0.12 8.4 2.5 12** 2.5 2.1 7.6

50% span length

0.32 1.9 1.4 3.5 1.2 2.5 2.3 6.3 4.4** 1.4 2.7* 4.3** 2.4 4.6 1.8 1.3 1.6 4.4 0.56 0.13 1.7 0.13 0.90 1.5 0.56 5.5** 3.9 0.06 2.0 2.8 1.3 0.51 0.89 3.9 0.25 1.8 2.1 0.18 1.9 1.7 0.31 1.3 2.6** 1.1 3.2* 2.7 1.9 1.8

2.5% span length

2.6* 0.34 1.7** 2.8*** 4.5*** 1.2 3.5*** 2.5** 1.0 1.0 1.3** 1.4 0.66 2.7 1.9 1.6 3.5*** 0.73 0.84 0.70 0.03 0.00 1.6 2.3 1.7 2.4* 3.3** 1.2 2.1 0.26 0.52 0.46 0.16 2.03 0.65 1.9 0.06 0.92 3.8*** 0.03 1.58 1.14 0.72 1.4 0.00 3.5** 0.22 0.66

Short ber content

11 26** 10 24** 8.0 22* 5.2 6.8 3.6 0.90 1.4 18* 14 3.4 7.8 6.5 5.1 3.6 6.1 0.40 10 7.9 5.6 12 19* 16* 3.9 14 24* 9.9 4.6 8.2 17* 23* 20* 21* 4.9 19 4.4 26* 4.7 15 5.7 6.7 1.8 17 8.1 8.8

Fineness
7.3*** 0.70 6.6*** 6.0*** 3.4* 5.6*** 1.7 2.3 1.1 3.4** 3.1** 2.1 3.0** 8.3*** 2.5** 5.9*** 8.5*** 5.5*** 6.2*** 8.0*** 6.4*** 5.9*** 1.1 7.3*** 2.8* 3.6** 0.21 1.7 1.4 3.0** 1.8 4.0** 3.0* 0.96 2.8** 2.3* 3.2** 2.5* 3.3** 0.81 2.4* 3.5* 4.4*** 0.35 1.2 0.64 0.46 0.17

*Signicant at p 0.05. **Signicant at p 0.01. ***Signicant at p 0.001.

multiple crosses must be conducted to combine all desirable traits in breeding programs designed to simultaneously improve lint yield and fiber quality. Regardless, positive
66

GCA effects in SP156 and SP224 for both lint yield and strength out of 48 F2 hybrids provides evidence for occurrence of recombination between these two unfavorably
CROP SCIENCE, VOL. 51, JANUARY FEBRUARY 2011

WWW.CROPS.ORG

associated traits in SP germplasm. The detection of positive GCA effects for lint yield, micronaire, and elongation in SP192, SP205, and JC65 indicates that these traits can be simultaneously improved.

of FM832 SP205 with 14% HPH and 48% MPH for lint yield had potential to be used in production for improving lint yield although a high yielding parent was not included in this study for estimation of useful heterosis.

Heterosis
Differences (p 0.001) between parent and hybrid (i.e., parent vs. hybrid) were observed for lint yield, strength, and span lengths (Table 1). Mean squares of parent vs. hybrid for these traits were at least 1.7 times larger than those of parents or hybrids. These results indicate significant heterosis in F2 hybrids for these traits and limited inbreeding depression since a 50% decrease of dominance from F1 generation to F2 generation is expected. Among the traits with significant differences between parent and hybrid, significant (parent vs. hybrid) environment interaction was observed only for lint yield and not for strength or span lengths (Table 1). Results indicate the relative stability of heterosis across environments for these fiber properties. High parent heterosis (HPH) for lint yield and fiber properties was estimated by comparing F2 hybrids to their high parent across four environments (Table 4). Six of the 48 F2 hybrids exhibited positive HPH for lint yield. MD15 and SP205 were the parents of five and two crosses, respectively, in the six hybrids. The F2 of FM832 SP205 yielded 14% more lint than the highest yielding parent, FM832. The remaining five F2 populations MD15 SP156, MD15 SP205, MD15 SP224, MD15 SP235, and MD15 JC65, yielded 12 to 26% more lint than their high parent, MD15. Six F2 hybrids (i.e., 12.5% of the 48 F2 hybrids), FM832 JC14, FM832 JC60, FM832 JC186, MD15 SP224, MD15 JC14, and MD15 JC60, exhibited favorable HPH in two of the seven fiber properties (Table 4). Thirteen F2 hybrids (27% of the 48 F2 hybrids) exhibited HPH in one of the seven fiber property. MD15 and FM832 as female parents and JC14 and JC60 as male parents were good combiners for strength (4.412% HPH) and span lengths (2.64.4% HPH for 50% span length and 3.5% HPH for 2.5% span length). A 20% HPH for elongation was identified in the F2 of TAM98D-99ne SP205. Tang et al. (1993b) detected positive HPH in 1.6, 3.1, 1.6, and 3.1% of 64 F2 hybrids for strength, micronaire, 2.5% span length, and 50% span length, respectively. In comparison, significant HPH was detected in 6.3, 10, 4.2, and 10% of 48 F2 hybrids in this study for the same properties, respectively. The higher percentages of significant HPH detected in this study may be attributed to the exotic genetic backgrounds of male parents in crosses. The majority of F2 hybrids were either equal to or exceeded mid-parent values for lint yield (83%), 50% span length (75%), and 2.5% span length (81%) (data not shown). Since skewed distributions over midparent values were observed among F2 hybrids for these traits, partial dominance would be a reasonable genetic base for the observed midparent heterosis (Fehr, 1987). F2 hybrids
CROP SCIENCE, VOL. 51, JANUARY FEBRUARY 2011

CONCLUSIONS
General combining ability effects (p 0.001) from exotic germplasm lines were identified for lint yield and fiber properties. General combining ability effects were more important than SCA effects for most traits except elongation and 50% span length. These results suggested that selection progress for lint yield and fiber properties could be expected in F2 hybrids derived from these exotic germplasm. Favorable GCA effects detected in all male parents for fiber properties provide evidence for the potential of the exotic germplasm in continuing genetic improvement of fiber quality in upland cotton. The crosses FM832 SP205 and MD15 SP205 produced a large amount of heterosis for lint yield that may be useful for promoting lint yield in cotton production using F2 hybrids. Acknowledgments
We thank Dr. Jack McCarty, Crop Science Research Laboratory in Mississippi State, MS, and Dr. Jixiang Wu, College of Agriculture and Biological Sciences, South Dakota State University, for reviewing manuscript. This research was founded by USDA-ARS, Project No. 6402-21000-033-00D.

References
Beil, G.M., and R.E. Atkins. 1967. Estimates of general and specific combining ability in F1 hybrids for grain yield and its components in grain sorghum, Sorghum vulgare Pers. Crop Sci. 7:225228. Constable, G.A., P.E. Ried, and N.J. Thomson. 2001. Approaches utilized in breeding and development of cotton cultivars in Australia. p. 116. In J.N. Jenkins and S. Saha (ed.) Genetic improvement of cotton: Emerging technologies. Science Publ. Inc., Enfield, NH. Cox, D.J., and K.J. Frey. 1984. Combining ability and the selection of parents for interspecific oat matings. Crop Sci. 24:963967. Coyle, G., and C.W. Smith. 1997. Combining ability for within-boll yield components in cotton, Gossypium hirsutum L. Crop Sci. 37:11181122. Culp, T.W., and D.C. Harrell. 1975. Influence of lint percentage, boll size, and seed size on lint yield of upland cotton with high fiber strength. Crop Sci. 15:741746. Culp, T.W., and D.C. Harrell. 1979. Registration of SC-1 cotton. Crop Sci. 19:410. El-Adl, A.M., and P.A. Miller. 1971. Transgressive segregation and the nature of gene action for yield in an intervarietal cross of upland cotton. Crop Sci. 11:381384. Fehr, W.R. 1987. Principles of cultivar development. Vol. 1. Theory and technique. Macmillan Publishing Company, New York, NY. Green, C.C., and T.W. Culp. 1990. Simultaneous improvement of yield, fiber quality, and yarn strength in upland cotton. Crop Sci. 30:6669. Hallauer, A.R., and J.B. Miranda. 1981. Quantitative genetics in maize breeding. Iowa State Univ. Press, Ames, IA. Jiang, C., P. Chee, X. Draye, P. Morrell, C.W. Smith, and A. Patterson. 2000. Multi-locus interactions restrict gene flow in

WWW.CROPS.ORG

67

advanced-generation interspecific populations of polyploidy Gossypium (cotton). Evolution Int. J. Org. Evolution 54:798814. Lee, J.A., P.A. Miller, and J.O. Rawlings. 1967. Interaction of combining ability effects with environments in diallel crosses of upland cotton (Gossypium hirsutum L.). Crop Sci. 7:477481. Liu, S., R.G. Cantrell, J.C. McCarty, Jr., and J.McD. Stewart. 2000. Simple sequence repeat- based assessment of genetic diversity in cotton race stock accessions. Crop Sci. 40:14591469. McCarty, J.C., J.N. Jenkins, W.L. Parrott, and R.G. Creech. 1979. The conversion of photoperiodic primitive race stocks of cotton to day-neutral stocks. Mississippi Agric. Forestry Exp. Stn. Res. Rep. 4(19):4. Meredith, W.R., Jr. 1984. Quantitative genetics. p. 131150. In R.J. Kohel and C.F. Lewis (ed.) Cotton. Agron. Monogr. 24. ASA, CSSA, and SSSA, Madison, WI. Meredith, W.R., Jr. 2005. Registration of MD52ne high fiber quality cotton germplasm and recurrent parent MD 90ne. Crop Sci. 45:807808. Meredith, W.R., Jr. 2006. Registration of MD15 upland cotton germplasm. Crop Sci. 46:2722. Meredith, W.R., Jr., and R.R. Bridge. 1972. Heterosis and gene action in cotton, Gossypium hirsutum L. Crop Sci. 12:304310. Meredith, W.R., Jr., and R.R. Bridge. 1973. The relationship between F2 and F3 progenies in cotton (Gossium hirsutum L.). Crop Sci. 13:354356. Meredith, W.R., Jr., and J.S. Brown. 1998. Heterosis and combining ability of cottons originating from different regions of the United States. J. Cotton Sci. 2:7784. Miller, P.A., and A. Marani. 1963. Heterosis and combining ability in diallel crosses of upland cotton, Gossypium hirsutum L. Crop Sci. 3:441444. Miller, P.A., and J.O. Rawlings. 1967. Breakup of initial linkage blocks through intermating in a cotton breeding program. Crop Sci. 7:199204. Percy, R.G., R.G. Cantrell, and J. Zhang. 2006. Genetic variation for agronomic and fiber properties in an introgressed recombinant inbred population of cotton. Crop Sci. 46:13111317. SAS Institute. 2004. SAS/STAT. Users guide. Version 9.1. SAS Inst.,

Cary, NC. Simmonds, N.W., and J. Smartt. 1999. Principles of crop improvement. Blackwell Science, Malden, MA. Smith, C.W., and G.G. Coyle. 1997. Association of fiber quality parameters and within-boll yield components in upland cotton. Crop Sci. 37:17751779. Sprague, G.F., and L.A. Tatum. 1942. General vs. specific combining ability in single crosses of corn. Agron. J. 34:923932. Tang, B., J.N. Jenkins, J.C. McCarty, and C.E. Watson. 1993a. F2 hybrids of host plant germplasm and cotton cultivars: I. Heterosis and combining ability for lint yield and yield components. Crop Sci. 33:700705. Tang, B., J.N. Jenkins, J.C. McCarty, and C.E. Watson. 1993b. F2 hybrids of host plant germplasm and cotton cultivars: II. Heterosis and combining ability for fiber properties. Crop Sci. 33:706710. Thaxton, P.M., C.W. Smith, and R. Cantrell. 2005. Registration of TAM 98D102 and TAM 98D99ne upland cotton germplasm lines with high fiber strength. Crop Sci. 45:16681669. Uster AFIS. 1977. Instruction manual of the advanced fiber information system: Instrument for measuring neps, length, diameter, and trash for fibers. Zellweger Uster, Knoxville, TN. Young, N.D., and S.D. Tanksley. 1989. RFLP analysis of the size of chromosomal segments retained around the tm-2 locus of tomato during backcross breeding. Theor. Appl. Genet. 77:353359. Zeng, L., and W.R. Meredith, Jr. 2009a. Associations among lint yield, yield components, and fiber properties in an introgressed population of cotton. Crop Sci. 49:16471654. Zeng, L., and W.R. Meredith, Jr. 2009b. Registration of five exotic germplasm lines of cotton derived from multiple crosses among Gossypium tetraploid species. J. Plant Reg. 3:7780. Zeng, L., W.R. Meredith, D.L. Boykin, and E. Taliercio. 2007. Evaluation of an exotic germplasm population derived from multiple crosses among Gossypium tetraploid species. J. Cotton Sci. 11:118127. Zeng, L., W.R. Meredith, Jr., and B.T. Campbell. 2010. Registration of four exotic germplasm lines derived from an introgressed population of cotton. J. Plant Reg. 4:240243.

68

WWW.CROPS.ORG

CROP SCIENCE, VOL. 51, JANUARY FEBRUARY 2011