Published online 31 May 2007
Published in Crop Sci 47:1013-1017 (2007)
© 2007 Crop Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
CROP BREEDING & GENETICS
Germplasm Potential for Trait Improvement in Upland Cotton: Diallel Analysis of Within-Boll Seed Yield Components
Paul Irwin Ragsdalea,* and
C. Wayne Smithb
a Defense Group Inc., 1140 Connecticut Ave. NW, Ste. 1140, Washington, DC 20036
b Dep. of Soil and Crop Sciences, Texas A&M Univ., College Station, TX 77843-2474. Research reported herein was conducted by the Texas Agric. Exp. Stn., the Texas A&M Univ. System
* Corresponding author (paul.ragsdale{at}defensegp.com).
 |
ABSTRACT
|
|---|
Within-boll seed yield components (seed per boll, motes per boll, ovules per boll, and seed-setting efficiency [SSE]) were evaluated in a diallel analysis of eight upland cotton (Gossypium hirsutum L.) genotypes. Genotypes included four converted race stock (CRS) accessions representing putative extremes for SSE and four commercial types representing Texas, mid-South, and eastern cotton production regions. Parents and their F1 progeny without reciprocals were evaluated in 2000 and 2001 near College Station, TX. General combining ability (GCA) effects and specific combining ability effects (P = 0.05) were found for each trait, indicating potential for improvement of seed yield components in this population. The CRS accession M-9044-0162 had the best mean performance and GCA effects for each trait, suggesting that it would be a good parent for improving within-boll seed yield components in upland cotton.
Abbreviations: CRS, converted race stock • GCA, general combining ability • M/B, motes per boll • O/B, ovules per boll • S/B, seed per boll • SCA, specific combining ability • SSE, seed-setting efficiency.
|
INTRODUCTION
|
|---|
CROP IMPROVEMENT requires the ability to select higher-performing individuals from segregating populations. In modern cotton (Gossypium hirsutum L.) breeding, this ability is limited by three factors. First, there is a lack of variation for traits of interest due to a narrow genetic base (Van Esbroeck et al., 1998, 1999). This results in a limited supply of alleles for traits of interest that can only be expanded by introducing (introgressing) novel alleles from other populations. The shift in cotton production to transgenic cultivars has exacerbated the narrowing of the genetic base of modern cottons and threatens to reduce the potential for genetic gain in lint yield (Bowman et al., 2003). Second, introgression from germplasm sources can be difficult due to incompatibility of related species or by photoperiodism within accessions of the same species. To facilitate their use in cotton breeding programs, 79 race stock accessions were converted to day neutrality for flowering (McCarty and Jenkins, 1993). The resulting converted race stocks (CRS) represent a potentially useful source of genetic variation. Third, essentially all wild sources of germplasm have undesirable agronomic traits that reduce their potential for elite cultivar development. In addition to known impairment of agronomic traits, Liu et al. (2000) found additional drawbacks of using the CRS germplasm population. First, there was variable recovery of exotic alleles during the conversion process that led to a loss of genetic variation from the original race stock source. The second drawback of the CRS population is that the original race stocks were heterogeneous and heterozygous, and this has been transmitted to their CRS progeny. This heterogeneity and heterozygosity complicates both genetic analyses and breeding selection progress.
The inheritance of and interrelationships among yield components is of interest to cotton breeders. The contribution to yield of various yield components was reviewed by Worley et al. (1974, 1976) and Heitholt (1999), who noted key contributions from the number of bolls per plant and lint percentage and that seed per boll (S/B) was the second largest contributor to lint yield after bolls per unit land area. Simultaneous improvement in multiple yield components provides an opportunity to increase yield using straightforward selection parameters. Cotton fibers are produced on the surface of seeds, so an increase in the number of seed of similar size should lead to an increase in fiber production. No study has yet determined the potential of exotic germplasm sources to improve within-boll seed yield components in elite cotton types. The objective of this research was to determine the potential to improve within-boll seed yield components by a diallel analysis of eight upland cotton genotypes.
|
MATERIALS AND METHODS
|
|---|
Four CRS accessions, two cultivars, and two germplasm lines of upland cotton were hand crossed in a diallel mating scheme at the Texas Agricultural Experiment Station near College Station, TX. Parents and their F1 progeny, without reciprocals, were evaluated in single-plant culture, 0.3 by 2.0 m, to minimize interplant competition. Experiments were conducted in 2000 and 2001 in a randomized complete block design with four replications of five plants each. Common cultural practices for cotton production in central Texas, including furrow irrigation, were used. Plots were direct seeded on 11 to 12 May 2000 and on 24 May 2001 in Belk clay, a fine, mixed, thermic Entic Hapludert.
Four CRS accessions were selected for the diallel study to represent putative extremes of seed-setting efficiency (SSE) based on preliminary results obtained from field studies conducted near College Station in 1998 (Nasirci and Smith, 1999). Accessions M-8844-0096 and M-9044-0162 have putatively high SSE; M-9044-0237 and M-9044-0244 have putatively low SSE. Two cultivars and two public germplasm lines were included in the diallel study to represent the diversity of adapted genotypes across production areas from Texas, the mid-South, and southeast USA. The cultivar TAM 94L-25 is a recent germplasm release of the Texas Agricultural Experiment Station (Smith, 2003). This strain has apparent drought tolerance, competitive yield and excellent fiber properties, large seed with relatively few S/B, and low SSE. The cultivar Tamcot CAMD-E is a near-obsolete, ultra-short-season cultivar released in 1979 by the Texas Agricultural Experiment Station (Bird, 1979). It has short fibers, poor fiber strength, low micronaire, and unknown SSE (Coyle and Smith, 1997; Nasirci and Smith, 1999). Deltapine Acala 90 is a mid-South-type cultivar released in 1981 by the Delta and Pine Land Company. It has small seed and unknown SSE (Coyle and Smith, 1997; Nasirci and Smith, 1999). The unreleased breeding line PD 6186 is a germplasm line made available in 1984 by the USDA-ARS in South Carolina. It has excellent fiber quality, large seed, and unknown SSE (Nasirci and Smith, 1999).
To help identify plants that harbor relatively more of the genetic variation from the original race stock source, individuals were selected within each CRS accession before serving as parents in the diallel mating scheme (Liu et al., 2000). Selection was conducted before flowering and was based on the presence of more primitive characteristics for plant height, multiple branches per node, and stem and leaf pubescence. Such individuals should harbor relatively more of the exotic parent genome than individuals that more closely resemble the commercial donor parent. To help alleviate heterogeneity within CRS accessions, plant-to-plant crosses were conducted to facilitate post-harvest selection based on expected SSE. Individual plant-to-plant combinations for which one or the other parent had aberrant SSE were discarded (data not shown). Reciprocal effects were ignored in this experiment, because little variation has been observed among upland cotton cytoplasms (Meredith, 1984) and because all CRS lines contain the same commercial cytoplasm type (McCarty and Jenkins, 1993).
At least two phenotypically normal, full-sized bolls that had opened naturally were hand harvested from each parent and F1 plant bearing fruit in 2000 and 2001. First- and second-position bolls from the middle of the fruiting zone were preferentially harvested to minimize effects due to fruiting position on the traits measured. Similarly, full-season bolls that had opened naturally were chosen to help minimize any effects due to late or variable maturity. Bolls were picked with the carpel walls (burs) intact to ensure recovery of all motes, i.e., failed reproductive structures that do not contribute to lint yield (Rea, 1929a, 1929b). Within-boll seed yield components of S/B, motes per boll (M/B), ovules per boll (O/B), and SSE were determined on a plot basis. Approximately equal numbers of bolls were included from each plant to help ensure balanced data for subsequent diallel analyses and to minimize sampling bias associated with plant-to-plant variation in the CRS accessions. A seed was defined as a full-sized seed that resisted crushing. A mote was defined as all other reproductive structures in the locule, ranging from small unfertilized ovules (approximately 1-mm diameter) to nearly full-sized seed structures that could be crushed easily (pops). However, it is unclear if these pops produce spinnable fibers and contribute to lint yield. Nevertheless, pops were rare in this population (data not shown), so any bias attributable to inclusion of pops in the mote tally should be minimal. The total number of ovules was determined by summing seed and motes. The SSE was determined by dividing S/B by O/B.
Parental data were analyzed using analysis of variance procedures with means separated by Tukey's control adjustment for multiple comparisons using Type III sums of squares (SAS Institute, 2002). Simultaneous confidence intervals for all pairwise comparisons among parental genotypes were generated using an experimentwide error rate of 0.05. Diallel analyses of variance were conducted using DIALLEL-SAS05 as described by Zhang et al. (2005) for Model I, Method 2 of Griffing (1956). The within-boll seed yield components data set had three missing observations out of 288, and trait values were predicted for these missing observations to facilitate use of DIALLEL-SAS05 that requires balanced data (Y. Zhang, personal communication, 2005). Prediction was used in favor of mean estimation, because one hybrid (M-8844-0096 x M-9044-0244) was present in only two plots in 2001. Combined with a significant genotype x year interaction that required mean estimates or predictions for each year separately, the resulting estimates for M-8844-0096 x M-9044-0244 in 2001 would have had only a single degree of freedom in 2001. Prediction of the missing observations allowed the use of genotypic data across years to determine more robust values for subsequent diallel analysis of variance. Trait values were predicted using the general linear model for Griffing's Model I, Method 2 described by Zhang and Kang (1997).
|
RESULTS AND DISCUSSION
|
|---|
Analysis of Parents
The eight parents differed for S/B, M/B, and SSE but not O/B (Table 1). Year effects were not significant for any trait, and genotypes responded differently to years only for O/B. Averaged across both years, M-9044-0162 had high S/B and SSE along with low M/B at 33.3, 0.89, and 4.2, respectively (Table 2). Genotype M-9044-0237 had low S/B and SSE along with high M/B at 27.2, 0.78, and 7.5, respectively. This range of variation demonstrates the rationale for selecting these parents for the diallel study. The frequency of motes in this study is higher than in some published reports (Davidonis et al., 2000; Saranga et al., 1998). This could be attributable to genotypic effects of the lines chosen for this population, but as environmental factors such as drought appear to have more substantial effects on mote number (Saranga et al., 1998), it appears that environmental effects due to location or season are a likely cause for this observation. There were no differences (P = 0.05) in S/B and SSE across the commercial-type checks in 2000 and 2001, which was unexpected because previous experience suggested that TAM 94L-25 would have lower S/B and SSE (Nasirci and Smith, 1999). As trends were apparent in the parental data, a larger sample size appears to be warranted for better separation of the genotypes to facilitate selection decisions. Nevertheless, these data suggest that M-9044-0162 has potential to improve within-boll seed yield components in this population.
View this table:
[in this window]
[in a new window]
|
Table 1. Analysis of variance of within-boll seed yield components for parental genotypes of cotton (Gossypium hirsutum L.) evaluated near College Station, TX, in 2000 and 2001.
|
|
View this table:
[in this window]
[in a new window]
|
Table 2. Mean within-boll seed yield components of eight parental genotypes of cotton (Gossypium hirsutum L.) when grown near College Station, TX, in 2000 and 2001.
|
|
Diallel Analysis
Diallel Analysis of Variance
Analysis of variance indicated an effect of years, genotypes, and an interaction of genotypes x years for all within-boll seed yield components for the eight parental genotypes and their 28 F1 hybrids (Table 3). Partitioning of the genotype and genotype x year interaction terms indicated general combining ability (GCA) and specific combining ability (SCA) effects (P < 0.001) and GCA x year effects (P < 0.05) for all within-boll seed yield components. The SCA x year effects (P < 0.01) were detected for all traits except O/B. These data indicate that there is sufficient genetic variation in this population to facilitate selection for each of these traits.
View this table:
[in this window]
[in a new window]
|
Table 3. Diallel analysis of variance for within-boll seed yield components of eight cotton (Gossypium hirsutum L.) genotypes and their F1 progeny when grown near College Station, TX, in 2000 and 2001.
|
|
General Combining Ability Effects
The GCA effects represent the average performance of parents in hybrid combinations, and thus which parents combined best with all other parents in the study. The diallel analyses of variance showed significant interactions of GCA x year for all traits, thus necessitating calculation of GCA effects for each year separately. Converted race stock M-9044-0162, which exhibited high S/B and low M/B (Table 2), produced F1 hybrids with the other seven parents in this study with an average increase of 0.9 and 1.8 S/B, an average decrease in M/B of 0.8 and 1.3, and an increase in SSE of 0.02 and 0.04 in 2000 and 2001, respectively (Table 4). No other CRS or commercial-type genotype was consistent in improving S/B, M/B, and SSE in both years. As expected, M-9044-0244 combined with the other seven parents to reduce S/B by 0.9 and 1.4 in 2000 and 2001, respectively. Genotype M-9044-0244 also increased M/B and decreased SSE in 2001.
View this table:
[in this window]
[in a new window]
|
Table 4. General combining ability effects for within-boll seed yield components from eight upland cotton (Gossypium hirsutum L.) genotypes when grown near College Station, TX, in 2000 and 2001.
|
|
Specific Combining Ability Effects
The SCA effects represent the deviation in hybrid performance from that expected from the GCA effects of each parent. Good specific combiners can be used to develop superior hybrids or to identify populations harboring complementary alleles from which superior genotypes can be selected. It should be noted that good SCA effects do not necessarily indicate superior trait performance. Two lines with poor overall trait performance can combine better than expected based on their GCA effects, but their hybrid progeny could still have poor trait performance. Therefore, combinations of lines that have beneficial SCA effects, as well as beneficial GCA effects and good-to-superior mean trait performance as parents per se, are of particular interest. In this population, one such combination was observed for SSE (Table 5). The combination of M-8844-0096 and M-9044-0162 demonstrated positive SCA effects for SSE of 0.04 in both years (P = 0.03 and 0.08 in 2000 and 2001, respectively), while possessing positive GCA effects for SSE and good-to-superior mean trait performance as parents per se.
View this table:
[in this window]
[in a new window]
|
Table 5. Specific combining ability effects for seed-setting efficiency among eight upland cotton (Gossypium hirsutum L.) genotypes when grown near College Station, TX, in 2000 and 2001.
|
|
|
CONCLUSIONS
|
|---|
It is clear that significant genotypic variation for within-boll seed yield components can be detected in the field. Moreover, the diallel analysis demonstrated sufficient genetic variation in these traits to facilitate trait improvement through selection, although the genotype x year interaction suggests that multiyear screening may be needed. On a practical basis, determination of S/B requires less time and effort than M/B or SSE so it seems that direct selection for S/B would be more efficient in a breeding scheme to improve these within-boll seed yield components and possibly yield through the contribution of S/B as a primary yield component so long as the number of bolls per unit land area is maintained.
The selected CRS accessions demonstrated useful variation for traits of interest in this population, suggesting their potential utility for trait improvement in commercial cotton populations. As is common for trait introgression from exotic sources, a backcross breeding approach probably will be necessary to introgress improved within-boll seed yield components into elite lines while minimizing impairment of other agronomic traits. Liu et al. (2000) suggested that marker-assisted selection should be conducted to maximize the probability of recovering exotic alleles in a backcross scheme. To ensure capturing of exotic alleles for within-boll seed yield components and to maintain the integrity of the commercial recurrent parent for agronomic traits, simultaneous selection could be conducted for within-boll seed yield components and for other highly heritable morphological traits. This simple phenotypic screening for traits to be introgressed (e.g., S/B) coupled with morphological selection for recurrent parent traits could provide a low-cost alternative to molecular marker-assisted selection for trait improvement in this population. To maximize recovery of exotic alleles and to reduce linkage drag associated with donor parent traits (e.g., S/B), it seems that large populations should be screened for within-boll seed yield components and morphological traits at each backcross stage.
The parent with the greatest potential to improve within-boll seed yield components in this population is CRS accession M-9044-0162. Consistently across both years, M-9044-0162 increased S/B, reduced M/B, and increased SSE. Additionally, this accession was found to perform unexpectedly well for SSE in combination with M-9044-0096, suggesting that the two accessions might possess complementary alleles for SSE and that a population derived from the two might be useful to increase SSE potential in upland cotton.
|
ACKNOWLEDGMENTS
|
|---|
We thank Dr. Manjit S. Kang and Dr. Yudong Zhang for their invaluable advice and counsel in the analysis of unbalanced diallel data sets with DIALLEL-SAS05. This research was funded in part by a USDA National Needs Fellowship in Plant Biotechnology, C. Everette Salyer Fellowship in Cotton Research, Tom Slick Graduate Research Fellowship, and Cotton Incorporated.
|
NOTES
|
|---|
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.
Received for publication September 29, 2006.
|
REFERENCES
|
|---|
- Bird, L.S. 1979. Registration of Tamcot CAMD-E cotton. Crop Sci. 19:411–412.[Free Full Text]
- Bowman, D.T., O.L. May, and J.B. Creech. 2003. Genetic uniformity of the U.S. upland cotton crop since the introduction of transgenic cottons. Crop Sci. 43:515–518.[Abstract/Free Full Text]
- Coyle, G.G., and C.W. Smith. 1997. Combining ability for within-boll yield components in cotton, Gossypium hirsutum L. Crop Sci. 37:1118–1122.[Abstract/Free Full Text]
- Davidonis, G.H., A. Johnson, and J.A. Landivar. 2000. Cotton mote frequency under rainfed and irrigated conditions. J. Cotton Sci. 4:1–9.
- Griffing, B. 1956. Concept of general and specific combining ability in relation to diallel crossing systems. Aust. J. Biol. Sci. 9:463–493.[Medline]
- Heitholt, J.J. 1999. Cotton: Factors associated with assimilation capacity, flower production, boll set, and yield. p. 235–269. In D.L. Smith and C. Hamel (ed.) Crop yield: Physiology and processes. Springer, Berlin.
- 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. 40:1459–1469.[Abstract/Free Full Text]
- McCarty, J.C., Jr., and J.N. Jenkins. 1993. Registration of 79 day-neutral primitive cotton germplasm lines. Crop Sci. 33:351.
- Meredith, W.R. 1984. Quantitative genetics. p. 131–150. In R.J. Kohel and C.F. Lewis (ed.) Cotton. ASA Monogr. 24. ASA, CSSA, and SSSA, Madison, WI.
- Nasirci, Z., and C.W. Smith. 1999. Characterization of 79 converted race stocks of upland cotton. p. 485–487. In Proc. Beltwide Cotton Conf., Orlando, FL. 3–7 Jan. 1999. Vol. 1. Natl. Cotton Council, Memphis, TN.
- Rea, H.E. 1929a. Varietal and seasonal variation of "motes" in upland cotton. J. Am. Soc. Agron. 21:481–486.
- Rea, H.E. 1929b. The influence of "motes" on the yield and boll-size of the cotton plant. J. Am. Soc. Agron. 21:1154–1155.
- Saranga, Y., N. Sass, Y. Tal, and R. Yucha. 1998. Drought conditions induce mote formation in interspecific cotton hybrids. Field Crops Res. 55:225–234.[CrossRef]
- SAS Institute. 2002. SAS release 9.00. SAS Inst., Cary, NC.
- Smith, C.W. 2003. Registration of TAM 94L-25 and TAM 94J-3 germplasm lines of upland cotton with improved fiber length. Crop Sci. 43:742–743.[Free Full Text]
- Van Esbroeck, G.A., D.T. Bowman, D.S. Calhoun, and O.L. May. 1998. Changes in the genetic diversity of cotton in the USA from 1970 to 1995. Crop Sci. 38:33–37.[Abstract/Free Full Text]
- Van Esbroeck, G.A., D.T. Bowman, O.L. May, and D.S. Calhoun. 1999. Genetic similarity indices for ancestral cotton cultivars and their impact on genetic diversity estimates of modern cultivars. Crop Sci. 39:323–328.[Abstract/Free Full Text]
- Worley, S., T.W. Culp, and D.C. Harrell. 1974. The relative contributions of yield components to lint yield of upland cotton, Gossypium hirsutum L. Euphytica 23:399–403.[CrossRef][ISI]
- Worley, S., H.H. Ramey, Jr., D.C. Harrell, and T.W. Culp. 1976. Ontogenetic model of cotton yield. Crop Sci. 16:30–34.[Abstract/Free Full Text]
- Zhang, Y., and M.S. Kang. 1997. DIALLEL-SAS: A SAS program for Griffing's diallel analyses. Agron. J. 89:176–182.[Abstract/Free Full Text]
- Zhang, Y., M.S. Kang, and K.R. Lamkey. 2005. DIALLEL-SAS05: A comprehensive program for Griffing's and Gardner–Eberhart analyses. Agron. J. 97:1097–1106.[Abstract/Free Full Text]
TOP
ABSTRACT
INTRODUCTION
