Cotton Breeding for Fiber Quality Improvement - Industrial Crops: Breeding for BioEnergy and Bioproducts - page 218

selection should be used as part of the breeding strategy to more effectively
assemble desired alleles for yield and fiber quality and to weaken and/or break
their negative relationship. Outliers such as in Fig. 10.5 with higher yield and fiber
strength need to be used in further intermating, with increased population sizes
during evaluation in subsequent generations. Bowman and Gutierrez [ 59 ] and May
[ 27 ] also recommended larger breeding population sizes in seeking improved fiber
quality and identifying transgressive segregants. Culp et al. [ 145 ] showed desirable
strength at a frequency of only 0.3 % in breeding populations, but they found the
frequency of desirable combinations increased in subsequent intermating.
Diallel studies have identified sources of improvement for fiber quality and yield
based on the combining ability of select breeding lines [ 146 , 147 ]. This breeding
analysis tool aids in determining the usefulness and variability of a program's
breeding material by qualifying and quantifying the trait of interest and its gene
action. Additive gene action has predominately been found more than nonadditive
for fiber quality especially when contrasting genotypes are used [ 148 - 150 ]; how-
ever, there have been reports of greater nonadditive gene action for fiber length
[ 151 - 153 ]. A review of gene action studies for lint yield and fiber quality are
reported in Meredith [ 154 ] and May [ 27 ]. As indicated in section “ Molecular
Approaches to Yield and Quality Enhancement ,” most fiber properties have been
found to have many QTL, confirming multigenic control and quantitative inheri-
tance which present challenges for using molecular markers.
Backcrossing may be used to transfer one or a limited number of desirable traits
from one parent (possibly not adapted) to an elite or adapted parent. Meredith [ 141 ]
proved fiber strength improvement was achievable when utilizing backcrossing;
however, it is commonly used at the early stages of transgenic trait development.
Stiller et al. [ 155 ] in a study of breeding methods with transgenic traits highlighted
that at least three backcrosses were required, but even then, there was considerable
variation in yield, fiber quality, and disease resistance in the progeny. They con-
cluded that a standard pedigree method was required to even recover the recurrent
parent characteristics. The aim of backcrossing should be to equal, if not exceed,
the recurrent parent performance.
In order to improve fiber quality for end users, breeders must know how
particular properties relate to spinning or spun yarn quality. In this sense, yarn
quality provides the fullest description of fiber quality. However, spinning yarn
from breeding populations is a high-cost exercise particularly if proper control of
spinning parameters (yarn count, twist, and production speed) is applied. Therefore
breeding programs use fiber testing instruments such as HVI, AFIS, and others as a
means for predicting yarn quality [ 156 - 161 ]. May and Green [ 152 ] reported that
selecting solely for fiber strength was not effective for improving yarn strength
indicating that there are more properties than fiber strength affecting yarn strength.
The value of assessing or even predicting yarn is that it provides information on the
integration of all fiber properties and their interactions in supporting yarn structure.
Breeders should be carefully measuring routinely selecting for improvement in
fiber properties while also selecting for yield and other key attributes. An HVI is
adequate for measuring fiber length and strength in breeding, despite being more