Abstract
The purpose of this selection study is to compare the use of fiber data in breeding programs.  Two selection methods, one using HVI data only and one using HVI plus AFIS data, will be compared.  The study begins with fifteen F₂ progenies that will undergo a series of selections through the F₄ generation.  In the final stages of the project, the remaining progeny lines will be used under go yarn and spinning quality analysis.  Fiber data from the second year of the study is presented and progress of the two selection methods is discussed.
Introduction

Traditionally cotton breeders have used HVI (High Volume Instrument) as their only source of fiber quality data when making plant selections.  This method of determining which plants to keep needs further evaluation, because the demands that breeders face and the expectations concerning fiber quality are intensifying.  Breeding programs must prove to be efficient and up to date if they are going to remain competitive as an integral part of the cotton industry.  As the cotton industry changes, breeders must think globally and continually strive to be proactive rather than reactive.  The purpose of this study is to provide data that will encourage the inclusion of AFIS in breeding programs, and demonstrate the value of such data.

The efficiency of selections made in breeding programs using two different methods will be compared.  One method uses only HVI data for progeny selections, and the second method uses both HVI and AFIS data for progeny selection.  The advantage of having AFIS data available is the additional information (including distributions) and fiber trait measurements.  Length measurements based on number as opposed to weight, short fiber content, fineness, and maturity ratios are all measurements that are offered by AFIS and not offered by HVI.  More specifically, AFIS provides length distributions that will be used as a critical part of the selection criteria for this test. There are data supporting the theory that AFIS may be an effective tool in predicting spinning performance and yarn quality (Hequet et al., 2007).  The ability to select for end performance would be a valuable tool for breeders.  F₂ progenies will be followed through a series of selections to the stage of “end product” by conducting spinning tests on progeny lines that are considered superior and selected in the F₄ generation.  The spinning performance and yarn quality of the chosen lines from each selection method will be compared.  A comparison of the progress made by each of the selection methods will be conducted in the final phase of the project.  This information as well as the fiber quality of the lines will be used as an indication of which selection method is more effective.  The final questions will be: 1. Did we accomplish more using both HVI & AFIS, compared to using HVI alone?, and 2.How well were we able to choose genotypes that performed well in spinning and the production of quality yarns?.  Breeders need to know that they can select genotypes that will be sought by the textile mills, and this research should offer hope that they can do just that if they have the appropriate resources.  The challenge of producing cotton that can compete in a world market must be met from several angles, and the quality of the lines found in breeding programs is an excellent place to start.

ObjectivesTo compare the efficiency of a breeding program using only HVI versus using HVI and AFIS data when making plant selections, starting with the F2 generation.  Selections will be made and lines will be evaluated through the spinning process in the final stages of the project.  In the final stages of the project, the progress of each selection method (one set of fiber data vs. two sets) will be compared.
Materials and Methods
The first year of the study, 2005, consisted of two separate, but identical (in design) field tests.  Fifteen F₂ populations and one commercial variety were planted in randomized complete blocks with four replications with two-row plots, and thinned to approximately one plant every eight inches.  Plot lengths were twenty-nine and one half feet (9 meters).  Bolls samples for family means obtained by picking one boll per plant.  Individual plant selections (IPS), sixteen per plot, were made based on agronomic performance and appearance.  The harvested samples, both boll samples and individual plants were ginned on a small, table-top, saw gin.  Check samples were ginned every fifty samples to evaluate the consistency of gin function.  Humidity and temperature readings were taken in the ginning area once every minute.  Gin performance was tested and environmental conditions were documented, because of the possible affects both could have on fiber quality results.  Fiber samples from one test underwent fiber analysis on both the Advanced Fiber Information System (AFIS) and High Volume Instrument (HVI).  The other set of fiber was only tested with HVI.  All fiber analysis was done by Texas Tech International Textile Center at Lubbock, TX.  Selections were made in two phases, first families, and then individual plants within the selected families.  Family selections were made based on the average fiber characteristics across replications.  Plant selections were based on key fiber characteristics of the individual plant in comparison to other plants (both within a family and across families).  The order of consideration for fiber properties is illustrated in Figure 1.  The test with HVI and AFIS data included length distributions as selection criteria in addition to the other AFIS data.  Each of the two tests was considered separately, resulting in two unique sets of selected plants for advancement.  Each selection method and the resulting entries will be considered a separate test for the duration of the trial.
In 2006, sixty-four F₃ populations and two commercial checks were planted in a randomized complete block designs with two replications and one-row plots.  Plots were thinned to approximately one plant every eight inches.  Replication number and plot size were small due to limited seed supply.  Each test consisted of only two replications with plot lengths of twenty feet (6.1 meters).  Again, boll samples were harvested for family means, and individual plants were selected.  More emphasis was put on the agronomic performance of plants; therefore there was not a predetermined amount of plants selected per row.  Ginning and fiber analysis was conducted using the same course of action as the previous year.  Fiber analysis differed some from the previous year.  All fiber samples from both tests underwent both HVI and AFIS analysis.  The AFIS data was needed for the HVI only test so comparisons could be made.  The AFIS data for the HVI only test was not used for selections; it was only used for comparisons after selections were made.  The selections based on fiber characteristics were done differently than the previous year, because of the caliber of fiber quality that was present.  For example, 2006 family values for strength averaged 36 g/tex for both tests with a minimum of 29.5.  The average upper half mean length (UHML) for  families in the HVI only test was 1.34.  The UHML average for HVI + AFIS test was 1.29 inches with the average length by number being .98 inches.  In 2006, there were not two distinct phases of selection.  Family means were of such high quality, no family was discarded based on family means alone.  Therefore, all individual plants from all families were considered for advancement.  At this point, plant selections were made by process of elimination.  The appropriate fiber characteristics (for each method) for individual plants were used for a series of selections.  Plants that performed below average in one or more traits were discarded until a manageable amount of lines were chosen for the 2007 test.  There were 151 plants selected for the HVI and AFIS method, and 153 plants chosen for the HVI only method.  
There are a total of four tests planted for 2007.  This includes two tests for the selection study (one for each selection method), and two yield tests (one for each selection method).  Each of the four tests is a randomized complete block design with four replications.  The yield tests have two-row plots, and the selection study tests has one-row plots.  At the time of this manuscript, these tests are in the field and coming along nicely.  Some tentative plans have been made for harvest, ginning and test analyses.
Results and Discussion

Two important questions that can be addressed with the current data available are,  1) Is there a difference in fiber quality between the two tests?, and 2) Have the selections been successful up to this point in terms of improving fiber quality?.  To illustrate apparent differences, distributions of fiber traits have been presented for both the subset of selected plants and the entire population of all plants prior to selection.  Chi-squared tests were used to identify statistical differences between the two test distributions for each fiber trait. Length & Related measurements (weight based)

One major difference between HVI and AFIS is that several measurements recorded by HVI are weight based, as where AFIS provides additional measurements that are based on number of fibers.  Weight based measurements provided by AFIS are upper quartile length by weight (UQLW) and mean length by weight (LW).  HVI provides upper half mean length (UHML) and length uniformity index (uniformity, UI).  The differences between these two measurement types are apparent in these data.  When looking at the length measurements based on weight (UQLW, LW, and UHML) in Figures 2 through 7, the HVI only test appears to have better plants in terms of length.  This difference is visible in both data sets (all plants, and selected plants), but is more obvious in the selected plants.  Chi-squared values (Table 1) indicate there are significant differences between the two tests, both for the original plant set and the selected plant set distribution, for the UHML, LW, and UQLW values.  The UHML distribution for the HVI only tests shows more change between the original plant set and the selected plants than the HVI + AFIS test (Figures 6 and 7).  This is not surprising because UHML was the only length measurement used for the HVI only selections.  The other two weight based measurements related to length are length uniformity and short fiber content (SFCW).  For these traits there are more plants with favorable values in HVI + AFIS test.  This test has more plants with lower SFCW, and fewer plants with a low uniformity index.  This is true when considering all plants (Figure10), but is more apparent after selections were made (Figure 11).  It is notable that in the distribution of selected plants there are no plants in the bottom three uniformity bins (Figure 9) or the upper six SFCW (Figure 11) bins for either test.  Plants within an undesirable range for these traits were eliminated.  Even more interesting, is that more plants were eliminated from a larger range in the HVI + AFIS test.  There are not any plants with SFCW values above 6.5 %, as where the HVI only test has plants in all bins up to 8.0 %.  Regarding uniformity, there are fewer plants in the HVI + AFIS test with values below 85.5 % than are in the HVI only test (Figure 9).  This is of interest because plants were not discarded solely because of either of these measurements.  Since these two traits were not individually considered as selection criteria, it is not expected that there be such an obvious difference between the original set of plants and the selected plants. (Note: plants with high SFCW and or low UI were eliminated because of poor fiber length distributions.)  This difference may be explained by the elimination of plants based on other traits like micronaire and or maturity depending on which test is being considered.  In the case of HVI + AFIS test, the difference between it and the HVI only test is probably the result of using addition information such as length distributions for selection criteria.  SFCW and uniformity were not considered individually for selection or rejection.  But, plants with low UHML values were rejected during the selection process.  It is evident that the lower limit varied between the two tests for the selected plants.  Once again this difference is due to consideration being given to length distributions. Length & Related measurements (number based)

Plant distributions for the other length and short fiber measurements based on number are different than the by weight measurements.  These length parameters, length by number (LN) and short fiber content by number (SFCN), show the HVI + AFIS test to be better regarding fiber length.  Once again this development is apparent in both the plant sets (Figure 12), but is more evident in the selected set (Figure 13).  For the selected plants, the dominant group in the HVI only test is the .98 in. bin with 16% of the plants, while the largest group in the HVI + AFIS test is the 1.00 in.  bin with 25% of its plants.  Another obvious difference between the two tests is the fact that there is a difference in the upper limit for SFCN in the selected plants (Figure 15); there are no plants above 21.5 %. There is also a visible difference in the lower limit for LN for the two tests.  The Chi-squared values (Table 1) were significant for both SFCN and LN for the selected plants; therefore there were differences for those two traits between tests.  These differences are the result of the two selection methods, since neither measurement was available in the HVI only test.  The original plant set was only significantly different between tests for the SFCN values.  This difference between the two tests before selections is an indication that the previous years’ selections had been successful.  The HVI + AFIS test having fewer plants with high SFCN supports the idea that this trait is heritable and can effectively be used as a selection criteria.  If this continues, the 2007-2008 population will have an even lower SFCN average.  It is critical to notice how different the performance of each test is depending on which group of measurements is considered (by weight versus by number).  Progress from previous selections is evident when comparing all plants for both tests, because of there being a difference in length properties between the two tests.  This year’s improvements are apparent when comparing the selected plants to the original set of plants. Strength and Elongation

The HVI measurements of strength and elongation were used as selection criteria for both tests.  Both tests were given a lower limit for strength, nothing under 31 g/tex was accepted.  This is apparent when comparing the entire plant population in Figure 16 to the selected population in Figure 17.  No definite patterns were seen with elongation values, other than the selected plants have fewer plants in the lower ranges (Figure19).  These two traits showed very little variability between tests.  For these two measurements, little discrepancy between the two tests is expected because the HVI measures them both and AFIS does not offer an individual measurement that is comparable.  Differences that occurred in the plant distribution were possibly a result of selecting for other traits such as fiber maturity.  A mature fiber generally is stronger than a similar fiber with less maturity.  Maturity and other related measurements may offer explanation for many of the observations previously discussed.  Fiber Maturity

One trait that is of particular interest, because of developments that occurred in these data, is fiber maturity.  Improving fiber maturity is especially important to the high plains cotton industry.  Many times cottons grown in this area have fiber properties would be appealing to a textile mill if it were not for the micronaire.  As seen in Figure 20, it is not uncommon for a large percentage of the cotton grown in this area to have low micronaire.  Improving fiber maturity, in addition to other fiber traits being superior, would give West Texas cotton a much needed competitive edge.  There were obvious differences between tests for several maturity related measurements.  In the beginning, micronaire was the sole trait used for the first round of eliminations for both tests.  Any plant with a micronaire value below 3.5 was discarded.  In the HVI only test, plants with a micronaire above 5.0 were also rejected.  In the HVI + AFIS test, some plants above 5.0 were selected based on other measurements.  In Figure 22, there are a higher percentage of selected plants with micronaire values in the 3.5-4.0 range in the HVI only test than remain in the HVI + AFIS test.  This range is not considered discount, but these values are low enough to indicate the possibility of fiber immaturity.  This becomes evident when looking at the maturity ratios for the selected plants (Figure 24) in the HVI only test.  There are a high percentage of plants that should not have been selected because of their maturity ratios.  It is important to remember, that for selection purposes micronaire was the only means of determining maturity for the HVI only test.  Micronaire is an indicator but not a direct measurement of maturity, therefore several of the plants with the same micronaire may have had different maturity ratios.  This gives reason for the lack of high maturity ratios in the HVI only test.  The most influential result that was seen in these data is an overall increase in maturity for the HVI + AFIS test.  Even before this years selection process the HVI + AFIS test had a higher percentage of plants with high maturity ratios (Figure 23) than the HVI only test.  This increase in maturity is not only coming from direct selection of high maturity ratios, but is also a result of selecting for good length (by number) distributions.  Plants with mature fibers tend to have better length distributions than plants with less mature fibers.  When discussing fiber maturities there are two other related fiber traits that should be considered, fineness and standard fineness. Fiber Fineness

Fineness values for all plants of both tests in Figure 25 do not have as clear a difference between the two tests as was seen with other traits.  However, chi-squared values in Table 1 showed the two tests to be significantly different.  The selected plant’s data (Figure 26) show that the HVI only tests has more plants with finer fibers. It is important to remember these fineness measurements may be skewed by differences in fiber maturity, which means these plants may not truly have finer fibers.  The fineness measurement is based on weight; therefore an immature fiber will have a lower fineness value.  Instead of having finer fibers, the HVI only test may have more plants with less mature fibers.  Maturity is also a likely explanation for why the lower fineness limit differs between all plants and the selected plants.  There are no selected plants with fineness values below 148 for either test; anything lower was indirectly eliminated based on micronaire or maturity values.  Fineness values were not used as an individual selection criterion and were not available at all for the HVI only selection process.  Because of the concerns with fineness values, it is important to also look at standard fineness. Standard fineness (Hs = fineness/maturity) gives and indication of fiber fineness while accounting for maturity.  In Figure 27 of this trait, there was no significant difference between the two tests for all plants.  However, there was a significant difference between the two tests for the selected plant set (Figure 28). Once again, the HVI only test appears to have more plants with finer fibers than the HVI + AFIS test.  Since maturity is accounted for with Hs, these low values may be accurate, and a result of selecting for other traits.  The longer and stronger the fiber is the greater tendency it will have to also be a finer fiber.  This relationship of traits most likely explains the distribution of fineness for the HVI only plants.  Because the main traits that were emphasized for selection in this test were length, strength, and micronaire it is probable that these plants would also have finer fibers.  Length Distribution

One selection criteria has been mentioned in relation to most all other fiber quality parameters and that is the length (by number) distribution.  Length by number distribution is made available by AFIS and played a very important role in the selection process for the HVI + AFIS test.  Plants were judged primarily on this fiber characteristic before much consideration was given to any other traits.  This was possible because so many fiber properties contribute to the shape of the distribution.  Figure 29 is a good example of what would be considered a good length distribution.  If a plant was lacking in any particular area such as length, strength, maturity, or amount of short fibers this weakness would become evident in the length distribution.  If a plant had a poor distribution it was discarded, regardless of other measurements.  Very rarely would a plant have a bad distribution, as illustrated in Figure 30, and still have any other desirable characteristics.  The length distribution might be described as a summary for the plants overall fiber quality.  Having this one characteristic to consider during the selection process is easier than considering multiple values for various traits simultaneously.  Had the length distribution been available for both tests there would likely be very little deviation in fiber quality between the two.  The length (by number) distribution is also thought to be useful in predicting the spinning performance of fibers.  This relationship will be tested in future stages of this study. 

Note: AFIS also provides a length by weight distribution.  However, it was not used because of problems related to weight based measurements that were mentioned previously.Conclusions

For now, we know that there is differentiation between the two tests for the two selection methods.  The majority of the fiber traits show both selection methods to be effective in eliminating plants with lesser fiber quality.  More importantly, the HVI + AFIS test is showing improvements for different traits than are seen in the HVI only test.  When looking at the plant population prior to selection for several traits it is visible that past selections have been successful.  Based on these data, it seems possible that by selecting for a desirable length distribution (by number) it was possible to indirectly improve fiber properties such as maturity and strength that directly impact the length distribution.  Differences between spinning performance are the greater interest, but cannot be determined yet.  However, these data have provided enough information for it to be suggested that there will be differences in spinning performance between the two tests.
References

1.      Baker, J.L., and L.M. Verhalen. 1973. The inheritance of several agronomic and fiber properties among selected lines of upland cotton, Gossypium hirsutm L. Crop Sci. 13: 444-450.

2. Benedict, C.R., R.J. Kohel, and H.L. Lewis. 1999. Cotton Fiber Quality. Pages 269-288 in C.W. Smith and J.T. Cothren, eds. Cotton: origin, history, technology, production. John Wiley & Sons, Inc. New York, NY.

3. Bowman, D.T. 2000. Contemporary Issues: Attributes of public and private cotton breeding programs. Journal of Cotton Science. 4:130-136.

4. Braden, C. A., and C.W. Smith. 2004. Fiber length development in near-long staple upland cotton. Crop Sci. 44:1553-1559.

5. Calhoun, S., and D.T. Bowman. 1999. Techniques for Development of New Cultivars. Pages 361-414 in C.W. Smith and J.T. Cothren, eds. Cotton: origin, history, technology, production. John Wiley & Sons, Inc., New York, NY

6. Culp, T.W. and D.C. Harrell. 1973. Breeding methods for improving yield and fiber quality of upland cotton (Gossypium hirstutum L.).Crop Sci.13:686-689.

7. Deussen, H. 1992. Improved cotton fiber properties – the textile industry’s key to success in global competition. P. 43-63. In  C.R. Benedict and G.M. Jividen (ed.) Cotton Fiber Cellulose: Structure, Function and Utilization Conference 1992. Natl. Cotton Council, Memphis, TN.

8. El-Zik, K.M., and P.M. Thaxton. Simultaneous improvement of yield, fiber quality traits, and resistance to pests of MAR cottons. P. 315 – 331. In  C.R. Benedict and G.M. Jividen (ed.) Cotton Fiber Cellulose: Structure, Function and Utilization Conference 1992. Natl. Cotton Council, Memphis, TN.

9. Green, C.C., and T.W. Culp. 1990. Simultaneous improvement of yield, fiber quality, and yarn strength, in upland cotton. Crop Sci. 30:66-69.

10. Latimer, S.K., T.P. Wallace, and D.S. Calhoun. 1996. Cotton breeding: High volume instrument versus conventional fiber quality testing. P. 1681. In P. Dugger and D.A. Richter (ed.) Proc. Belt. Cotton Res. Conf., New Orleans, LA. 9-12 Jan. 1996. Natl. Cotton Council, Memphis, TN.

11. May, O.L. and G.M. Jividen. 1999. Genetic modification of cotton fiber properties as measured by single- and high-volume instruments. Crop Sci. 39:328-333.

12. May, O.L., and R.A. Taylor. 1998. Breeding cottons with higher yarn tenacity. Textile Res. J. 68:302-307.

13. Meredith, W.R. , Jr., T.W. Culp, K.Q. Robert, G.F. Ruppenicker, W.S. Anthony, and J.R. Williford. 1991. Determining future cotton variety fiber quality objectives. Textile Res. J. 61:715-720.

14.  Meredith, W.R. Jr., P.E. Sasser, and S.T. Rayburn. Regional high quality fiber properties as measured by conventional and AFIS methods. P 1681-1684. In P. Dugger and D.A. Richter (ed.) Proc. Belt. Cotton Res. Conf., New Orleans, LA.  9-12 Jan. 1996.  Natl. Cotton Council, Memphis, TN.

15.   Price, J.B., T.A. Calamari, and W.Y. Tao. 1999. Yarn Preparation, Fabric Formation, and Finishing. Pages 751-791. in C.W. Smith and J.T. Cothren, eds. Cotton: origin, history, technology, production. John Wiley & Sons, Inc., New York, NY.

		All Plants		Selected Plants
		Chi Squared Test		Chi Squared Test
Mic		0.0754		0.0000
UHML		0.0000		0.0000
Uniformity		0.0313		0.0025
Strength		0.2139		0.8368
Elongation		0.1116		0.7277
Length by weight		0.0016		0.0003
UQLW		0.0000		0.0000
Short Fiber Content (w)		0.0006		0.0000
Length by number		0.5583		0.0143
Short Fiber Content (n)		0.0000		0.0000
Fineness		0.0146		0.0000
Maturity		0.2089		0.0000
Standard Fineness (Hs)		0.2016		0.0076

Table 1. Chi squared values for comparison of histograms.
Selection Process
HVI Only
Micronaire
(plants in the discount ranges were discarded)
Upper Half Mean Length
(only accepted lengths above a set minimum)
Strength
(made sure all plants met minimum strength requirements)
Uniformity
(made all plants were within an acceptable range)
HVI & AFIS
Length by number Distribution
(selected good distributions^* for further consideration)
Length by number
(only accepted lengths above a set minimum)
HVI Properties
(made sure all were within an acceptable range)

A “good” distribution is illustrated in Figure #.

Figure 1. A flow chart of what fiber properties were used for selection and the order in which they were considered for each selection method.

Figure 2. Histograms of upper quartile length by weight values for all individual plants (selected for agronomic traits in the field) in the HVI only test and the HVI + AFIS test.

Figure 3. Histograms of upper quartile length by weight values for selected plants (selected for fiber quality) in the HVI only test and the HVI + AFIS test.

Figure 4. Histograms of length by weight values for all individual plants (selected for agronomic traits in the field) in the HVI only test and the HVI + AFIS test.

Figure 5. Histograms of length by weight values for selected plants (selected for fiber quality) in the HVI only test and the HVI + AFIS test.

Figure 6. Histograms of upper half mean length values for all individual plants (selected for agronomic traits in the field) in the HVI only test and the HVI + AFIS test.

Figure 7. Histograms of upper half mean length values for selected plants (selected for fiber quality) in the HVI only test and the HVI + AFIS test.

Figure 8. Histograms of uniformity index values for all individual plants (selected for agronomic traits in the field) in the HVI only test and the HVI + AFIS test.

Figure 9. Histograms of uniformity index values for selected plants (selected for fiber quality) in the HVI only test and the HVI + AFIS test.

Figure 10. Histograms of short fiber content by weight values for all individual plants (selected for agronomic traits in the field) in the HVI only test and the HVI + AFIS test.

Figure 11. Histograms of short fiber content by weight values for selected plants (selected for fiber quality) in the HVI only test and the HVI + AFIS test.

Figure 12. Histograms of length by number values for all individual plants (selected for agronomic traits in the field) in the HVI only test and the HVI + AFIS test.

Figure 13. Histograms of length by number values for selected plants (selected for fiber quality) in the HVI only test and the HVI + AFIS test.

Figure 14. Histograms of short fiber content by number values for all individual plants (selected for agronomic traits in the field) in the HVI only test and the HVI + AFIS test.

Figure 15. Histograms of short fiber content by number values for selected plants (selected for fiber quality) in the HVI only test and the HVI + AFIS test.

Figure 16. Histograms of strength values for all individual plants (selected for agronomic traits in the field) in the HVI only test and the HVI + AFIS test.

Figure 17. Histograms of strength values for selected plants (selected for fiber quality) in the HVI only test and the HVI + AFIS test.

Figure 18. Histograms of elongation values for all individual plants (selected for agronomic traits in the field) in the HVI only test and the HVI + AFIS test.

Figure 19. Histograms of elongation values for selected plants (selected for fiber quality) in the HVI only test and the HVI + AFIS test.

Figure 20.  Micronaire values from the Lubbock classing office.  Percentage of total bales classed for a given micronaire.

Figure 21. Histograms of micronaire values for all individual plants (selected for agronomic traits in the field) in the HVI only test and the HVI + AFIS test.

Figure 22. Histograms of micronaire values for selected plants (selected for fiber quality) in the HVI only test and the HVI + AFIS test.

Figure 23. Histograms of maturity ratios for all individual plants (selected for agronomic traits in the field) in the HVI only test and the HVI + AFIS test.

Figure 24. Histograms of maturity ratios for selected plants (selected for fiber quality) in the HVI only test and the HVI + AFIS test.

Figure 25. Histograms of fineness values for all individual plants (selected for agronomic traits in the field) in the HVI only test and the HVI + AFIS test.

Figure 26. Histograms of fineness values for selected plants (selected for fiber quality) in the HVI only test and the HVI + AFIS test.

Figure 27. Histograms of standard fineness values for all individual plants (selected for agronomic traits in the field) in the HVI only test and the HVI + AFIS test.

Figure 28. Histograms of standard fineness values for selected plants (selected for fiber quality) in the HVI only test and the HVI + AFIS test.
Figure 29. Example of what is considered to be a good length (by number) distribution.

Figure 30. Example of what would be considered a bad length (by number) distribution.
Figure 31.  Average length distribution for the individual plants selected for each test.

		All Plants		Selected Plants
		Chi Squared Test		Chi Squared Test
Mic		0.0754		0.0000
UHML		0.0000		0.0000
Uniformity		0.0313		0.0025
Strength		0.2139		0.8368
Elongation		0.1116		0.7277
Length by weight		0.0016		0.0003
UQLW		0.0000		0.0000
Short Fiber Content (w)		0.0006		0.0000
Length by number		0.5583		0.0143
Short Fiber Content (n)		0.0000		0.0000
Fineness		0.0146		0.0000
Maturity		0.2089		0.0000
Standard Fineness (Hs)		0.2016		0.0076

Table 1. Chi squared values for comparison of histograms.
References

16.  Baker, J.L., and L.M. Verhalen. 1973. The inheritance of several agronomic and fiber properties among selected lines of upland cotton, Gossypium hirsutm L. Crop Sci. 13: 444-450.

17. Benedict, C.R., R.J. Kohel, and H.L. Lewis. 1999. Cotton Fiber Quality. Pages 269-288 in C.W. Smith and J.T. Cothren, eds. Cotton: origin, history, technology, production. John Wiley & Sons, Inc. New York, NY.

18. Bowman, D.T. 2000. Contemporary Issues: Attributes of public and private cotton breeding programs. Journal of Cotton Science. 4:130-136.

19. Braden, C. A., and C.W. Smith. 2004. Fiber length development in near-long staple upland cotton. Crop Sci. 44:1553-1559.

20. Calhoun, S., and D.T. Bowman. 1999. Techniques for Development of New Cultivars. Pages 361-414 in C.W. Smith and J.T. Cothren, eds. Cotton: origin, history, technology, production. John Wiley & Sons, Inc., New York, NY

21. Culp, T.W. and D.C. Harrell. 1973. Breeding methods for improving yield and fiber quality of upland cotton (Gossypium hirstutum L.).Crop Sci.13:686-689.

22. Deussen, H. 1992. Improved cotton fiber properties – the textile industry’s key to success in global competition. P. 43-63. In  C.R. Benedict and G.M. Jividen (ed.) Cotton Fiber Cellulose: Structure, Function and Utilization Conference 1992. Natl. Cotton Council, Memphis, TN.

23. El-Zik, K.M., and P.M. Thaxton. Simultaneous improvement of yield, fiber quality traits, and resistance to pests of MAR cottons. P. 315 – 331. In  C.R. Benedict and G.M. Jividen (ed.) Cotton Fiber Cellulose: Structure, Function and Utilization Conference 1992. Natl. Cotton Council, Memphis, TN.

24. Green, C.C., and T.W. Culp. 1990. Simultaneous improvement of yield, fiber quality, and yarn strength, in upland cotton. Crop Sci. 30:66-69.

25. Latimer, S.K., T.P. Wallace, and D.S. Calhoun. 1996. Cotton breeding: High volume instrument versus conventional fiber quality testing. P. 1681. In P. Dugger and D.A. Richter (ed.) Proc. Belt. Cotton Res. Conf., New Orleans, LA. 9-12 Jan. 1996. Natl. Cotton Council, Memphis, TN.

26. May, O.L. and G.M. Jividen. 1999. Genetic modification of cotton fiber properties as measured by single- and high-volume instruments. Crop Sci. 39:328-333.

27. May, O.L., and R.A. Taylor. 1998. Breeding cottons with higher yarn tenacity. Textile Res. J. 68:302-307.

28. Meredith, W.R. , Jr., T.W. Culp, K.Q. Robert, G.F. Ruppenicker, W.S. Anthony, and J.R. Williford. 1991. Determining future cotton variety fiber quality objectives. Textile Res. J. 61:715-720.

29.  Meredith, W.R. Jr., P.E. Sasser, and S.T. Rayburn. Regional high quality fiber properties as measured by conventional and AFIS methods. P 1681-1684. In P. Dugger and D.A. Richter (ed.) Proc. Belt. Cotton Res. Conf., New Orleans, LA.  9-12 Jan. 1996.  Natl. Cotton Council, Memphis, TN.

30.   Price, J.B., T.A. Calamari, and W.Y. Tao. 1999. Yarn Preparation, Fabric Formation, and Finishing. Pages 751-791. in C.W. Smith and J.T. Cothren, eds. Cotton: origin, history, technology, production. John Wiley & Sons, Inc., New York, NY.