Crop Science Grow Your Career with CSSA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 8 September 2006
Published in Crop Sci 46:2076-2080 (2006)
© 2006 Crop Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bednarz, C. W.
Right arrow Articles by Brown, S. M.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Bednarz, C. W.
Right arrow Articles by Brown, S. M.
Agricola
Right arrow Articles by Bednarz, C. W.
Right arrow Articles by Brown, S. M.
Related Collections
Right arrow Crop Growth and Development
Right arrow Cotton

CROP ECOLOGY, MANAGEMENT & QUALITY

Plant Density Modifications of Cotton Within-Boll Yield Components

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 Parkway, Cary, NC 27513
c Univ. of Georgia, Rural Development Center, P.O. Box 1209, Tifton, GA 31793

* Corresponding author (craig.bednarz{at}ttu.edu)


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
One approach to improving cotton (Gossypium hirsutum L.) yield and quality is to identify crop management practices that may exploit the most basic (i.e., within-boll) yield components. One of the parameters that may influence within-boll yield components is plant density. Thus, the objectives of this investigation were to determine how yield components in cotton are altered through plant density management. Two cotton cultivars were overseeded and hand thinned to 3.6, 9.0, 12.6, and 21.5 plants m–2 in 2001 and 2002. 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 yield component calculations. Lint mass boll–1, individual seed mass, and seed number boll–1 decreased as plant density increased while total seed surface area m–2 of land area increased, which resulted in increased lint yield m–2 of land area. Lint mass cm–2 of seed surface area and fiber number seed–1 did not consistently respond to plant density. These results indicate that plant density management may influence total seed surface area per unit land area. Most within-boll yield components, however, appear to be controlled more by cultivar than crop management.

Abbreviations: AFIS, Advanced Fiber Information System


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
THE RELATIONSHIPS among cotton 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 seed per boll while maintaining a high lint percentage (lint mass relative to seed cotton mass). These authors suggested more seed 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 successful Delta cultivars. Miller and Rawlings (1967) also found as yield increased by selection, lint percentage and seed per boll increased while boll and seed size decreased.

These findings illustrate within-boll yield components have changed as a result of selection for increased lint yield. The question arises, how are within-boll yield components altered through crop management and the environment? Current research has shown boll number per unit land area and boll and seed size are readily influenced by plant density (Bednarz et al., 2000). Thus, it seems reasonable that within-boll yield components are influenced by plant density as well. If this is true, is it possible to identify crop management practices that may capitalize on the most basic yield components? The objectives of this investigation, therefore, were to determine how yield components in cotton are altered through plant density management.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Plot Establishment and Maintenance
Experiments were conducted in 2001 at the Coastal Plain Experiment Station in Tifton, GA, on a Tifton loamy sand (fine-loamy, kaolinitic, thermic Plinthic Kandiudults) and in 2001 and 2002 at the Southwest Branch Experiment Station in Plains, GA, on a Greenville sandy clay loam (Fine, kaolinitic, thermic Rhodic Kandiudults). In March of each year, 672 kg ha–1 of 3–9–18 (N–P–K) plus micronutrients (8% Ca, 2% Mg, 9% S, 0.13% B, 0.10% Fe, 1% Mn, and 0.35% Zn) was broadcast and harrow incorporated. Trifluralin ({alpha},{alpha},{alpha}-trifluoro-2,6-dinitro-N,N-dipropyl-p-toluidine; 1.12 kg a.i. ha–1) was then broadcast and harrow incorporated immediately before ripping and bedding. While ripping and bedding, 32 kg a.i. ha–1 of 1,3-dichloropropene was injected under the row for nematode control at Tifton. All other fertility, weed, and insect-pest control practices 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) when soil water tension fell below –40 kPa at the 20-cm soil depth or below –50 kPa at the 40-cm soil depth. Cotton (‘DPL 458 BR’ and ‘FM 966’) was overseeded on 10 and 14 May 2001 at Tifton and Plains, respectively, and 12 May 2002 at Plains, and hand thinned to 3.6, 9.0, 12.6, and 21.5 plants m–2 on stand establishment (i.e., {approx}3 wk after planting). The cultivars chosen for this study were selected on the basis of seed size and were both managed as nontransgenic. Each plot was four rows (0.9-m centers) wide and 38 m long. Harvest aids were applied at {approx}70% open boll in each study, and were a combination of tribufos (S,S,S-tributyl phosphorotrithioate; 0.321 kg a.i. ha–1) plus thidiazuron (N-phenyl-N'-1,2,3-thiadiazol-5-ylurea; 0.093 kg a.i. ha–1) plus ethephon (2-chloroethyl phosphonic acid; 1.103 kg a.i. ha–1).

Data Collection
Immediately before machine harvest, {approx}1 m of row was removed from each end of all plots with a 2.1-m-wide mower (Bush Hog, Inc., Selma, AL) to avoid inaccurate measures of plot yield and fiber quality due to the end-of-row effect (Holman and Bednarz, 2001). Also 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, seed number, 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 Advanced Fiber Information System (AFIS) instrument (Uster Technologies, Charlotte, NC). 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 seed–1 was determined. Fiber fineness and lint mass seed–1 data were used to determine total fiber length seed–1. Mean fiber length data from 3000 individual fiber length measurements, also reported by AFIS, and total fiber length seed–1 data were then used to determine fiber number seed–1. 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 seed–1 and seed surface area data were then used to determine fiber number cm–2 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 between cultivar and plant density with location–year (Tifton, GA, 2001; Plains, GA, 2001; and Plains, GA, 2002). Thus, the data are presented for each location–year (Steel and Torrie, 1980). The data for each location–year were analyzed as a split plot in time using Proc MIXED (SAS Institute, 2000) where the main plots consisted of replications and eight treatments (two cultivars and four plant densities) arranged as a two by four factorial, and the split plots consisted of the main stem nodes (i.e., fruiting positions). After analysis, weighted means or sums were found by averaging or summing over the nodes to obtain data for the eight treatments.


   RESULTS AND DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Harrell and Culp (1976) suggested breeders should select for increased number of bolls per unit land area with more seed per boll. Previous studies (Bednarz et al., 2000) have indicated bolls per unit land area are readily influenced by plant density. The current investigation concurs as bolls m–2 increased with increasing plant density in both cultivars (Table 1). Bridge et al. (1971) and Miller and Rawlings (1967) indicated selection for increased lint yield resulted in cultivars with more seed per boll and smaller boll and seed sizes. The rationale for the increased number of small seed per boll was increased seed surface area within the boll for fiber development (Harrell and Culp, 1976). Our data indicate boll and seed mass and seed number per boll all decreased as plant density increased (Tables 1 and 2). Reduced seed number per boll in the current investigation with increased plant density was offset, however, by increased boll number per unit land area, resulting in increased seed number per unit land area (Table 2). While reduced surface area per seed also occurred as a result of increased plant density (Table 3), the end result was increased total seed surface area per unit land area (Table 3), which resulted in increased lint yield (Table 1).


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 plant density studies conducted at the University of Georgia in 2001 and 2002. The data presented are either the sum or mean of all fruiting positions.{dagger}

 

View this table:
[in this window]
[in a new window]
  		Table 2. Seed numbers per boll and per unit ground area and individual seed mass in plant density studies conducted at the University of Georgia in 2001 and 2002. The data presented are either the sum or mean of all fruiting positions.{dagger}

 

View this table:
[in this window]
[in a new window]
  		Table 3. Seed surface areas per seed and per unit ground area and lint mass per seed in plant density studies conducted at the University of Georgia in 2001 and 2002. The data presented are either the sum or mean of all fruiting positions.{dagger}

 
Arguably, seed size is the primary force driving yield components. Fiber Max 966 is a large-seeded cultivar while DPL 458 BR is small-seeded (Table 2). It is interesting to note, therefore, that yield components in this investigation differed greatly between the two cultivars. The larger-seeded cultivar 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 seed with less total seed surface area on a land area basis (Tables 1GoGo–4.). Cultivar differences with respect to lint yield, however, did not consistently differ across the three environments (Table 1).


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 plant density studies conducted at the University of Georgia in 2001 and 2002. The data presented are either the sum or mean of all fruiting positions.{dagger}

 
Two additional points regarding Tables 1Go–3 should be noted. First, some in the U.S. cotton industry have suggested that the migration to smaller-seeded cultivars through years of selection has resulted in cotton plants with reduced resilience in the presence of stress (primarily water deficit or high-temperature stress). It is hypothesized that smaller-seeded cultivars produce more seed per unit land area and are investing too much energy into seed production (Lewis, 2001). Since cotton seed is oil based and lint is cellulose based, more energy is required to produce the additional seed. Thus, in the presence of water deficit or high temperature stress, the crop has little energy for maintenance respiration and even less energy for growth or reproduction. While cultivar differences in total seed number per unit land area were observed in the current investigation, cultivar differences in total seed mass per unit land area were not observed. Total seed mass (g m–2) was as follows for the three environments and did not differ between cultivars: Plains 2001, DPL 458 BR = 168.48, FM 966 = 161.32; Tifton 2001, DPL 458 BR = 183.70, FM 966 = 177.87; Plains 2002, DPL 458 BR = 205.81, FM 966 = 185.34. These data suggest the energy requirement for seed oil production was roughly equivalent between the two cultivars.

On the contrary, it can also be argued that small-seeded cultivars are more adaptable to environmental stresses. While they did not offer an explanation, Culp and Harrell (1975) also suggested cultivars with medium to small bolls might adjust more rapidly to adverse environmental conditions. If smaller-seeded cultivars produce smaller but more bolls per unit land area, then each boll arguably requires less energy to produce. Thus, abortion of a boll would result in less wasted energy. Also, when the stress was released (e.g., through irrigation), less energy would be required to replace previously aborted bolls.

The second additional point regarding Tables 1Go–3 is the apparent changes in yield components with time. 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 {approx}30 to 39%. Boll and seed size ranged from {approx}6 to 8 g seed cotton boll–1 and from {approx}120 to 140 mg seed–1, respectively. The number of seed boll–1 averaged about 36 while lint seed–1 averaged about 72 mg for medium to small-seeded cultivars and 85 to 90 mg for the large-seeded cultivars. Finally, boll number m–2 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 40 to 43% across cultivars and years (data not presented). Boll and seed size across years was 1.8 and 2.5 g lint boll–1 and 69 and 94 mg seed–1 for DPL 458 BR and FM 966, respectively. The number of seed boll–1 across years was 31.5 and 32.3 for DPL 458 BR and FM 966, respectively, while lint seed–1 across years was 56 mg and 77 mg for DPL 458 BR and FM 966, respectively. Boll number was 88 and 63 m–2 for DPL 458 BR and FM 966, respectively. 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 and fewer seed boll–1. 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 seed boll–1 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) seed boll–1. 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). As indicated earlier, cultivar differences with respect to lint yield did not consistently differ across the three environments (Table 1). While the smaller-seeded cultivar (DPL 458 BR) 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). Thus, yields were roughly equivalent across environments. Also, fiber quality data from this study (Bednarz et al., 2005) indicated fineness of FM 966 was lower (i.e., finer) than DPL 458 BR. Interestingly, FM 966 produced about 3300 more fibers cm–2 than DPL 458 BR. Thus, the finer fiber produced by FM 966 may have resulted from these additional fibers produced per unit of seed surface area. These data indicate fiber number and lint mass per unit seed surface area are confounded with seed size, which should be considered when selecting for increased lint frequency.


   CONCLUSIONS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Interestingly, a large-seeded cultivar by current standards would have been considered small-seeded thirty years ago. It appears years of cultivar development for increased lint percentage has also resulted in reduced seed and boll sizes. Miller and Rawlings (1967) suggested increased lint yield through selection for increased lint percentage has not only reduced boll and seed size, 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 size. It has been suggested that 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 seed compensate for production of small bolls with less mass of fibers per seed through production of more bolls and seed 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 seed.

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 or decreased seed size, 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 criteria. If lint mass were increased 2 mg cm–2 of seed surface area in the two cultivars used in this investigation, lint yield would increase by {approx}45 kg ha–1 for each. Our data, however, indicate fiber number and lint mass per unit seed surface area are confounded with seed size, which should be considered when selecting for increased lint frequency.

Finally, in the cotton industry, seed are commonly sold in 250 000 seed count bags. In this investigation, DPL 458 BR (the smaller-seeded cultivar) produced {approx}750 more seed m–2 than did FM 966. This equates to 30 additional bags of cotton seed ha–1 for the sales inventory. All else being equal, why not produce smaller-seeded cotton cultivars?


   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
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
J. Article No. T-4-570.

Received for publication December 22, 2005.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bednarz, C. W.
Right arrow Articles by Brown, S. M.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Bednarz, C. W.
Right arrow Articles by Brown, S. M.
Agricola
Right arrow Articles by Bednarz, C. W.
Right arrow Articles by Brown, S. M.
Related Collections
Right arrow Crop Growth and Development
Right arrow Cotton

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Vadose Zone Journal
Journal of Plant Registrations Soil Science Society of America Journal
Journal of Natural Resources
and Life Sciences Education		 Journal of
Environmental Quality