Published online 1 September 2007
Published in Crop Sci 47:2108-2112 (2007)
© 2007 Crop Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
CROP ECOLOGY, MANAGEMENT & QUALITY
Within-Boll Yield Components of High Yielding Cotton Cultivars
Craig W. Bednarza,*,
Robert L. Nicholsb and
Steve M. Brownc
a Texas Tech Univ. and the Texas Agricultural Experiment Station, Box 42122, Lubbock, TX 79409
b Cotton Incorporated, 6399 Weston Pkwy., Cary, NC 27513
c Univ. of Georgia, Rural Development Center, P.O. Box 1209, Tifton, GA 31793. J. Article No. T-4-573
* Corresponding author (craig.bednarz{at}ttu.edu).
 |
ABSTRACT
|
|---|
Cotton (Gossypium hirsutum L.) within-boll yield components have changed throughout the last 30 yr of cultivar development. The question arises, how do within-boll yield components differ in contemporary high yielding cultivars? Nine commercially available cotton cultivars were over-seeded and hand-thinned to 10.8 plants m–2 in 2001, 2002, and 2003. Before machine harvest, plants from 6 m of one row were removed from each plot and hand-harvested by fruiting position. After hand-harvest, seed cotton from each fruiting position was ginned separately. Boll number, lint mass, seed number, seed mass, seed surface area, and fiber properties were determined for each fruiting position. These data were then used for within-boll yield component calculations. One of the top yielding cultivars in this investigation (DPL 33 B), characterized by a smaller seed mass, produced greater total seed surface area per unit of land area, but lower lint mass and fiber number per unit of seed surface area. The other two top yielding cultivars in this investigation (DPL 491 and STV 4892 BR), characterized by a larger seed mass, produced lower total seed surface area per unit of land area, but greater lint mass and fiber number per unit of seed surface area. These data indicate fiber number and lint mass per unit of seed surface area are linked to seed size, which should be considered when selecting for increased lint mass or fiber number per unit of seed surface area.
Abbreviations: AFIS, Advanced Fiber Information System • DD15, degree days with a base of 15°C • DPL, Delta and Pine Land • FM, Fibermax • PM, Paymaster • PSC, Phytogen Seed Company • STV, Stoneville
Within-Boll Yield Components of High Yielding Cotton Cultivars
Craig W. Bednarza,*,
Robert L. Nicholsb and
Steve M. Brownc
a Texas Tech Univ. and the Texas Agricultural Experiment Station, Box 42122, Lubbock, TX 79409
b Cotton Incorporated, 6399 Weston Pkwy., Cary, NC 27513
c Univ. of Georgia, Rural Development Center, P.O. Box 1209, Tifton, GA 31793. J. Article No. T-4-573
* Corresponding author (craig.bednarz{at}ttu.edu).
Cotton (Gossypium hirsutum L.) within-boll yield components have changed throughout the last 30 yr of cultivar development. The question arises, how do within-boll yield components differ in contemporary high yielding cultivars? Nine commercially available cotton cultivars were over-seeded and hand-thinned to 10.8 plants m–2 in 2001, 2002, and 2003. Before machine harvest, plants from 6 m of one row were removed from each plot and hand-harvested by fruiting position. After hand-harvest, seed cotton from each fruiting position was ginned separately. Boll number, lint mass, seed number, seed mass, seed surface area, and fiber properties were determined for each fruiting position. These data were then used for within-boll yield component calculations. One of the top yielding cultivars in this investigation (DPL 33 B), characterized by a smaller seed mass, produced greater total seed surface area per unit of land area, but lower lint mass and fiber number per unit of seed surface area. The other two top yielding cultivars in this investigation (DPL 491 and STV 4892 BR), characterized by a larger seed mass, produced lower total seed surface area per unit of land area, but greater lint mass and fiber number per unit of seed surface area. These data indicate fiber number and lint mass per unit of seed surface area are linked to seed size, which should be considered when selecting for increased lint mass or fiber number per unit of seed surface area.
Abbreviations: AFIS, Advanced Fiber Information System • DD15, degree days with a base of 15°C • DPL, Delta and Pine Land • FM, Fibermax • PM, Paymaster • PSC, Phytogen Seed Company • STV, Stoneville
|
INTRODUCTION
|
|---|
THE RELATIONSHIPS among cotton (Gossypium hirsutum L.) lint yield and its components are complex (Worley et al., 1974). Worley et al. (1974) concluded that boll number per unit land area was the largest contributor to lint yield, followed by seed number per boll and lint mass per seed. Culp and Harrell (1975) reported increased lint yield was possible under selection for medium to small bolls with the greatest possible number of small seeds per boll while maintaining a high lint percentage (lint mass relative to seed cotton mass). These authors suggested more seeds per boll may be desirable because of the greater amount of surface area for lint production within the boll (Harrell and Culp, 1976). Bridge et al. (1971) discovered a general change to smaller bolls, smaller seed and higher lint percentage in high yielding Delta cultivars. Miller and Rawlings (1967) also found as yield increased by selection, lint percentage and seeds per boll increased while boll and seed mass decreased. Data from recent field experiments in Georgia showed two contemporary cultivars possessed lower boll and seed mass than any found in the cotton literature (Bednarz et al., 2006).
These findings illustrate within-boll yield components have changed as a result of selection for increased lint yield. The questions arise, how are within-boll yield components related to yield in contemporary high yielding cotton cultivars? Do these characteristics differ from earlier reports? Are small boll and seed mass a common characteristic of contemporary high yielding cultivars? If it is possible to identify common within-boll yield components among high yielding cultivars, selection criteria for future cultivar development may be identified that could capitalize on these most basic yield components. Hence, the objectives of this investigation were to determine how yield components differ in contemporary high yielding cotton cultivars.
|
MATERIALS AND METHODS
|
|---|
Cultural Practices
Studies were conducted at the University of Georgia Coastal Plain Experiment Station Ponder Farm in 2001, 2002, and 2003 on a Tifton loamy sand (fine-loamy, kaolinitic, thermic Plinthic Kandiudult). Nine commercially available cotton cultivars were over-seeded on 9 May 2001, 18 June 2002, and 29 Apr. 2003 with a Monosem air planter (Lenexa, KS) on 91-cm row widths. The cotton cultivars utilized in this study and their respective maturity classifications as provided by their respective planting-seed purveyors were Phytogen Seed Company (PSC) 355 (early), Paymaster (PM) 1199R (early), Delta and Pine Land (DPL) 491 (early-mid), Fibermax (FM) 966 (mid-full), Stoneville (STV) 4892BR (mid-full), DPL 33B (mid-full), DPL Pearl (full), DPL 565 (full), and PSC GA161 (full). While planting, 6.7 kg ha–1 Aldicarb [2-methyl-2-(methylthio) propionaldehyde O-methylcarbamoyloxime] was applied in furrow for insect control. After emergence (approximately 14 d after planting) all plots were hand thinned to 10.8 plants m–2. Fertility, weed control, and insect scouting and control measures were in accordance with the University of Georgia Cooperative Extension Service guidelines (Brown et al., 2001). Water stress was minimized with overhead sprinkler irrigation in all studies. Irrigation water was applied (2.54 cm) to all plots when soil water tension at any monitoring location fell below –40 kPa at the 20 cm soil depth or below –50 kPa at the 40-cm soil depth. Harvest aids were applied (2.3 L ha–1 of ethephon plus cyclanilide and 0.7 kg ai ha–1 of thidiazuron) when the crop achieved 90% open boll (20 Sept. 2001, 13 Nov. 2002, and 16 Sept. 2003). The experimental design used was a randomized block design with four (2001 and 2003) or three (2002) replicates. Each plot was four rows (0.91-m spacing) wide and 15 m long.
Data Collection
Immediately before machine harvest, plants from 6 m of one of the center rows were removed from each plot and taken to a field laboratory where they were later hand-harvested by fruiting position. After hand-harvest, seed cotton from each fruiting position was ginned separately. Boll number, lint mass and seed number, and seed mass and seed surface area (after acid delinting) were determined for each fruiting position. Fiber from each fruiting position was delivered to Cotton Incorporated (Cary, NC) for fiber quality analyses. Fiber quality was determined using an Uster Technologies (Charlotte, NC) Advanced Fiber Information System (AFIS) instrument. Fiber fineness and mean fiber lengths from the AFIS data set were used to calculate fiber numbers. Fiber fineness is reported as fiber mass per unit fiber length (i.e., mg km–1). Also, through ginning and acid delinting, lint mass per seed was determined. Fiber fineness and lint mass per seed data were used to determine total fiber length per seed. Mean fiber length data from 3000 individual fiber length measurements, also reported by AFIS, and total fiber length per seed data were then used to determine fiber number per seed. Seed surface area was determined through alcohol displacement and a series of coefficients described by Hodson (1920). For this analysis, 100 seed were placed in a graduated cylinder and covered with 20 mL of ethanol. The volume of ethanol displaced was then used for determination of seed surface area using the coefficients developed by Hodson (1920). Fiber number per seed and seed surface area data were then used to determine fiber number per centimeter of seed surface area. Data presented in this manuscript are either the sum or weighted mean of all fruiting positions.
Statistical Analyses
The initial combined data analyses showed interactions with year. Thus, the data are presented for each year (Steel and Torrie, 1980). The data for each year were analyzed as a split plot in time using Proc MIXED (SAS Institute, 2000) where the main plots consisted of replications and nine cultivars, and the split plots consisted of the main stem nodes. After analysis, weighted means or sums were found by averaging or summing over the nodes to obtain data for the nine cultivars.
|
RESULTS
|
|---|
Because of the wide range in planting date among years and its possible effect on yield components between years, degree days (heat units with a base of 15°C [DD15]) were calculated. In 2001 the number of days elapsed from planting to defoliation was 135 with a total DD15 accumulation of 1391 degree days (°C d). In 2002 the number of days elapsed from planting to defoliation was 149 with a total DD15 accumulation of 1393°C d. Thus, the later planting date in 2002 resulted in two additional weeks to reach a roughly equivalent number of degree days. In 2003 the number of days elapsed from planting to defoliation was 140 with a total DD15 accumulation of 1394°C d.
The cultivars utilized in this investigation were high yielding Georgia cultivars, the recurrent parent of a high yielding Georgia cultivar, or were presented by the seed companies as prime candidates for potential to commercialize. Thus, this collection of cultivars represented the products of commercial breeding that were most highly regarded by the planting seed industry. Across years, the top three cultivars in terms of lint mass (g m–2) were DPL 33 B, DPL 491, and STV 4892 BR. While these cultivars were roughly equivalent in agronomic performance (i.e., lint yield), yield components among them clearly differed. Thus, our results will focus on these three cultivars and data from the other cultivars used in this investigation will not be presented. It would be informative, however, to survey yield components of lower yielding cultivars as well. An attempt was made during protocol development to select commercially available cultivars that were genetically diverse as well as high yielding. Unfortunately, agronomic performance of several of the selected cultivars was unacceptable, which was likely due to several factors other than yield components such as lack of ability to exploit a long season environment while tolerating intermittent periods of heat and drought stress. Thus, yield components of the underperforming (i.e., low yielding) cultivars will not be presented because it is suggested their unacceptable agronomic performance was likely due to their inability to acclimate to the Lower Coastal Plain and interrelations among yield components under these environmental conditions many not reflect the relationships under more optimal conditions.
Harrell and Culp (1976) suggested breeders should select for increased number of bolls per unit land area with more seeds per boll. Bridge et al. (1971) and Miller and Rawlings (1967) indicated selection for increased lint yield resulted in cultivars with more seeds per boll and smaller boll and seed sizes. Of the top three yielding cultivars, DPL 33 B produced the greatest number of bolls per square meter followed by STV 4892 BR and then DPL 491 (Table 1
). While DPL 33 B produced the greatest number of bolls per square meter, it also produced the least lint mass per boll in all years (Table 1). In 2002 and 2003, DPL 491 and STV 4892 BR were equally greater in lint mass per boll than DPL 33 B. In two of the 3 yr, DPL 33 B and DPL 491 produced the greatest number of seeds per boll while STV 4892 BR produced the least (Table 2
). Conversely, in all 3 yr, seed mass (mg seed–1) was lowest for DPL 33 B (Table 2). In summary, increased boll number per unit land area with more seeds per boll and smaller boll and seed sizes were suggested as desirable selection criteria more than 30 yr ago. DPL 33 B more closely matched these criteria than the other top yielding cultivars utilized in this investigation.
View this table:
[in this window]
[in a new window]
|
Table 1. Boll number per unit ground area and lint mass per boll and per unit ground area in studies conducted at the University of Georgia in 2001, 2002, and 2003.
|
|
View this table:
[in this window]
[in a new window]
|
Table 2. Seed number per boll and per unit ground area and individual seed mass in studies conducted at the University of Georgia in 2001, 2002, and 2003.
|
|
Harrell and Culp (1976) indicated the rationale for the increased number of small seeds per boll was increased seed surface area within the boll for fiber development. As previously mentioned, DPL 33 B produced more bolls per unit land area with more seeds per boll (Tables 1 and 2). This also resulted in a greater number of seeds produced per unit land area with this cultivar (Table 2). The lower seed mass (mg seed–1) of DPL 33 B (Table 2) was concomitant with lower seed surface area (cm2 seed–1). The additional number of seeds produced with this cultivar per unit land area, however, resulted in a total seed surface area per unit land area (m2 m–2) that was greater than any of the other cultivars included in the study (Table 3
). Thus, while we did not measure total seed surface area within the boll as suggested by Harrell and Culp (1976), total seed surface area per unit land area, which is arguably one of the most basic components of yield potential, was consistently greater in one (DPL 33 B) of the top three yielding cultivars in this investigation.
View this table:
[in this window]
[in a new window]
|
Table 3. Seed surface area per seed and per unit ground area and lint mass per seed in studies conducted at the University of Georgia in 2001, 2002, and 2003.
|
|
While the other two top yielding cultivars included in this investigation (DPL 491 and STV 4892 BR) produced less total seed surface area per unit land area, lint mass produced per seed (mg seed–1) was greater (Table 3). Greater lint mass per seed occurred through production of more fibers per seed with these two cultivars relative to DPL 33 B (Table 4
). In 2001 and 2002 the number of fibers produced per square centimeter of seed surface area and lint mass per square centimeter of seed surface area was greater in DPL 491 and STV 4892 BR relative to DPL 33 B (Table 4). Thus, while DPL 491 and STV 4892 BR produced less total seed surface area, high lint yield per unit land area was maintained through greater lint production per seed and per unit seed surface area.
View this table:
[in this window]
[in a new window]
|
Table 4. Fiber numbers per seed and per unit seed surface area (SSA) and lint mass per unit seed surface area in studies conducted at the University of Georgia in 2001, 2002, and 2003.
|
|
|
DISCUSSION
|
|---|
While we did not test this hypothesis, it is possible that seed size is the primary force driving yield components. DPL 491 and STV 4892 BR are larger seeded cultivars while DPL 33 B is small seeded (Table 2). It is interesting to note, therefore, that yield components in this investigation differed greatly between the cultivars. The larger seeded cultivars in this investigation produced fewer bolls per unit land area, more lint mass per boll, more lint mass and fiber number per seed, and fewer seeds with less total seed surface area on a land area basis (Tables 1–4
). Cultivar differences with respect to lint yield, however, did not consistently differ across the three environments (Table 1).
Coyle and Smith (1997) indicated the difficulty associated with their measurement resulted in little selection pressure for within-boll yield components other than lint percentage and concomitantly seed size. Harrell and Culp (1976) indicated high lint percentage would continue as the key selection criteria until a method to rapidly and economically determine lint frequency (i.e., lint mass per unit of seed surface area) was available. Culp and Harrell (1975) and Harrell and Culp (1976) conducted yield component studies using commercial cultivars and Pee Dee lines with release dates from 1945 to 1975. During this 30-yr period, lint percentage increased from approximately 30 to 39%. Boll and seed size ranged from approximately 6 to 8 g seed cotton boll–1 and from approximately 120 to 140 mg seed–1, respectively. The number of seeds per boll averaged about 36 while lint mass per seed averaged about 72 mg for medium to small seeded cultivars and 85 to 90 mg for the large seeded cultivars. Finally boll number per square meter for the commercial cultivars and Pee Dee lines released throughout this 30-yr period ranged from about 30 to 55. In the current investigation, lint percentage ranged from 38 to 41% across cultivars and years (data not presented). Boll and seed size across cultivars and years was 2.07 g lint boll–1 and 77.7 mg seed–1. The number of seeds per boll across cultivars and years was 29.0 while lint mass per seed across cultivars and years was 68.5. Boll number across cultivars and years was 72.8 m–2. Thus, it appears selection for increased yields during the last 30 yr has resulted in cultivars with increased lint percentage but smaller seed and boll masses with fewer seeds per boll and more bolls per square meter. As reported earlier, Bridge et al. (1971) and Miller and Rawlings (1967) indicated selection for increased lint yield resulted in cultivars during the late 1960s and early 1970s with more seeds per boll and smaller boll and seed sizes with more seed surface area within the boll for fiber development (Harrell and Culp, 1976). Modern cultivars, however, may be characterized by even smaller seed and boll masses, but also fewer (not more) seeds per boll. While this may have resulted in less seed surface area within the boll for fiber development, the additional number of bolls on a land area basis combined with smaller seed size may have resulted in a total seed surface area on a land area basis that is even greater today.
As early as 1920 Hodson (1920) suggested breeders should use lint frequency (i.e., lint mass per unit of seed surface area) to improve yield potential. While one single yield component cannot be considered alone, the weight of lint and number of fibers per unit seed surface area must be the most basic within-boll yield components (Coyle and Smith, 1997). While the smaller seeded cultivar (DPL 33 B) produced more total seed surface area on a land area basis (Table 3), lint mass on a seed surface area basis was lower (Table 4). The other two top yielding cultivars in this investigation (DPL 491 and STV 4892 BR), however, produced less total seed surface area on a land area basis but more lint mass on a seed surface area. These data indicate fiber number and lint mass per unit seed surface area are linked to seed size, which should be considered when selecting for increased lint frequency.
|
CONCLUSIONS
|
|---|
Interestingly, a large seeded cultivar by current standards would have been considered small seeded 30 yr ago. It appears years of cultivar development for increased lint percentage has also resulted in reduced seed and boll mass. Miller and Rawlings (1967) suggested increased lint yield through selection for increased lint percentage has not only reduced boll and seed mass, but also fiber length and fiber strength. Stewart and Kerr (1974) also indicated selecting for lint percentage alone to increase yield could compromise fiber length and seed mass. It has been suggested fiber length varies by fiber location on the seed, seed location within the boll, and boll location on the plant (Bradow and Davidonis, 2000). Our data suggest cultivars with small seeds compensate for production of small bolls with less mass of fibers per seed through production of more bolls and seeds per unit land area. Thus, if a particular location on the seed or within the boll is a source of short fibers, the problem could become exacerbated when inadvertently selecting for small seeds.
Throughout the last 60 yr of cultivar development it appears lint percentage has increased by as much as 10%. The question arises, how much more can lint percentage increase? If fiber quality becomes less desirable with increased lint percentage and decreased seed mass then further increases in lint percentage are not advisable. Thus, selection for increased lint mass per unit seed surface area may be the next reasonable selection criterion. Across cultivars and years, if lint mass were increased 2 mg cm–2 of seed surface area, lint yield would increase by approximately 35 kg ha–1.
|
ACKNOWLEDGMENTS
|
|---|
The authors would like to thank Benjamin G. Mullinix, Jr. for assistance with the statistical analyses and T. Dudley Cook and Lola C. Sexton for the technical support. The authors also thank the Georgia Agricultural Commodity Commission for Cotton and Cotton Incorporated for the financial support.
|
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 December 22, 2006.
|
REFERENCES
|
|---|
- Bednarz, C.W., R.L. Nichols, and S.M. Brown. 2006. Plant density modifications of cotton within-boll yield components. Crop Sci. 46:950–956.[Abstract/Free Full Text]
- Bradow, J.M., and G.H. Davidonis. 2000. Quantitation of fiber quality and the cotton production-processing interface: A physiologist's perspective. J. Cotton Sci. 4:34–64.
- Bridge, R.R., W.R. Meredith, and J.F. Chism. 1971. Comparative performance of obsolete varieties and current varieties of upland cotton. Crop Sci. 11:29–32.[Abstract/Free Full Text]
- Brown, S.M., S. Culpepper, G. Harris, P. Jost, B. Kemerait, P. Roberts, and D. Shurley. 2001. 2001 Georgia cotton production guide. Univ. of Georgia, Coop. Ext. Serv., Tifton.
- 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]
- 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. 15:741–746.[Abstract/Free Full Text]
- Harrell, D.C., and T.W. Culp. 1976. Effects of yield components on lint yield of upland cotton with high fiber strength. Crop Sci. 16:205–208.[Abstract/Free Full Text]
- Hodson, E.A. 1920. Lint frequency in cotton with a method for determination., Bulletin no. 168. Univ. of Arkansas, Fayetteville.
- Miller, P.A., and J.O. Rawlings. 1967. Selection for increased lint yield and correlated responses in upland cotton, Gossypium hirsutum L. Crop Sci. 7:637–640.[Abstract/Free Full Text]
- SAS Institute. 2000. SAS/C OnlineDoc. SAS Inst., Cary, NC.
- Steel, R.G.D., and J.H. Torrie. 1980. Principles and procedures of statistics: A biometrical approach. 2nd ed. McGraw-Hill, New York.
- Stewart, J.M., and T. Kerr. 1974. Relationship between fiber length increase and seed volume increase in cotton (Gossypium hirsutum L.). Crop Sci. 14:118–120.[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][Web of Science]
TOP
ABSTRACT
INTRODUCTION
