Wednesday, September 12, 2007 - 2:45 PM
Evidence on the possibility to develop cotton source germplasm based on the F1 cultivar cross and F2 and BC1 generation performance data
TITLE:
DISCIPLINE:
AUTHORS:
ACKNOWLEDGEMENT:
COVER PAGE Breeding & Genetics Panagiotis Michalakopoulos (Corresponding Author) Lab. of Plant Breeding and Biometry, Agricultural University of Athens, Greece. 118 55 Phone: +30 24220 21717 Fax: +30 24220 21717 E-mail : p6445674@otenet.gr C. Goulas Lab. of Forest Genetic and Breeding, Aristotelian Univ. of Thessaloniki, Greece. A. Katsiotis Lab. of Plant Breeding and Biometry, Agricultural University of Athens, Greece. 118 55. S.R. Sree Rangasamy Coimbatore, India. 641 041. To ICAC which selected the first author as a Research Associate to participate in the World Cotton Research Conference-4 to be held in Lubbock, Texas, USA from September 10-14, 2007.
Michalakopoulos
P. A.1, C. K. Goulas2, A. Katsiotis1, & S.R. Sree Rangasamy3 1Lab.
of Plant Breeding and Biometry, Agricultural University of Athens,
Greece. 2Lab.
of Forest Genetic and Breeding, Aristotelian Univ. of Thessaloniki,
Greece. 3
Agri R & D Consultant, Coimbatore, India.
ABSTRACT
The current cotton (G. hirsutum)
cultivars are of high yielding potential and early maturity combined
with good fiber quality traits. Breeders face the challenge to
improve yielding potential, both as lint and fiber quality taking
into account the great difficulties in improving fiber traits while
maintaining yield and fiber quality. The possibility to develop
genetic variability from six crosses using as parents adapted and
introduced cotton commercial cultivars to be utilized in applied
variety development program was the aim of this study. The six
generations (P1, P2,
F1, F2, BCP1, and BCP2)
from each F1 cross
were field evaluated in 2005. A randomized
complete block design (RCB) arrangement with three replications was
used. The analysis of variance for seed cotton yield produced
significant generation effects in four out of the six crosses
evaluated. Data provided evidence on effective differentiation among
some of the generations evaluated, such as the BCP1 in
cross GR1 x GR3 and BCP1, BCP2 in cross GR1 x
GR5, whereas significant mid-parent F1 and F2
heterosis effects were observed. Regarding earliness significant
generation effects in five out of the six crosses evaluated were
observed. On the contrary non-significant F1 and F2
mid-parent heterosis were observed. Regarding the fiber quality
traits, significant generation effects associated with F1
mid-parent heterosis were observed in four out the six crosses for
fiber length. The same response was observed for the traits:
micronaire, strength and uniformity in three, two and one cross
respectively. The data provided preliminary evidence on the value of
the particular segregating populations studied to be useful germplasm
sources in applied variety development programs.
Key words: backcrosses - cotton (G.
hirsutum) - F1 - F2 - heterosis. Introduction
Cotton, grown primarily for its lint, is a major
world crop grown in more than 60 countries, and is mainly planted,
about 96%, to Upland G. hirsutum (Meredith, 1999).
Current cotton cultivars have high yielding potential and early
maturity combined with good fiber quality traits. In spite of this,
cotton yields worldwide ending the 20th century seem to be
on a plateau in all cotton producing countries except for India and
Turkey (Chaudhry, 1997a). Thus, breeders face the challenge to
improve yielding potential, both as lint and fiber quality, taking
into account the great difficulties in improving fiber traits while
maintaining yield and fiber quality (Worley et al. 1976;
Calhoun and Bowman, 1999). The majority of
cotton breeders in developing varieties have used conventional
pedigree selection or some variation of this method, like modified
pedigree or modified bulk pedigree, as summarized by Calhoun and
Bowman (1999). The main difference among selection methods is
the managing of the segregating population and especially the
effectiveness of the early generation evaluation. Thus, selection
among rather than within early generation populations seems to be
more effective. Furthermore, pedigree selection has
been effective in improving quality traits such as micronaire and
fiber length, lint percent, seedcotton yield, lint yield and
earliness in appropriate populations in spite of the fact that these
traits express low heritability (Calhoun and Bowman, 1999).
The multi-adversity resistance (MAR) breeding approach aiming to
improve host plant resistance to diseases and other stresses while
maintaining or increasing yield is worth mentioning (Calhoun and
Bowman, 1999, and references herein). Cotton cultivars are developed by pedigree
selection methodology but rarely if ever are developed as pure lines
as in small grains or soybean (Pelhman and Sleper, 1995). This
indicates that both heterogeneity and heterozygosity are present in
cultivars and was capitalized in developing derived varieties through
selection within existing cultivars (Calhoun and Bowman, 1999).
Backcross breeding has been also used to transfer morphological
traits into various genetic backgrounds for genetic studies and
breeding lines development and to transfer major genes for fiber
strength (Meredith, 1993). This, of course, does not rule out use of
backcross for transferring quantitative traits, as well.
Cotton hybrid variety breeding has been employed
since both F1 and F2 useful heterosis has been
reported as summarized by Meredith(1999). In India F1
hybrids are widely grown (Chaudhry 1997b). No hybrid cotton
cultivars were commercialized in USA by the end of the 20th
century although the advent of commercialized transgenic traits may
make hybrid cotton cultivars more attractive (Calhoun and Bowman,
1999).
No matter the type of varieties, lines or
hybrids, to be developed, the choice of germplasm sources (parental
material) to form the genetic variability necessary for selection to
be practiced is crucial. Obsolete cultivars, introduced germplasm or
lines from other species and race stocks, developed by germplasm
enhancement programs are potential sources, but the first choice is
existing cultivars. Such commercial cultivars have an accepted
agronomic performance and usually some or few less desirable traits
that must be eliminated or improved. Such traits that are of prime
commercial breeder’s concern are yield, fiber quality and
appropriate maturity (Calhoun and Bowman, 1999). The same
holds true even for hybrid variety development since the
highest-yielding hybrids generally are derived by crossing the
highest-yielding parents (Davis, 1978).
High seedcotton yielding ability, early maturing and good fiber
quality traits are needed for cultivars to become commercially
successful under Greek cropping conditions. A wide range of
proprietary line cultivars, including early to late maturing
cultivars, having variable high seedcotton yielding ability and fiber
quality traits are currently used by farmers. Therefore research to
investigate the possibility to develop new germplasm using successful
cultivars as promising source material in applied breeding programs
was initiated.
To develop genetic variability from six crosses using as parents
adapted and introduced cotton commercial cultivars to be utilized in
applied variety development program was the objective of this study.
Materials and Methods
Seven proprietary commercial line cotton cultivars
were evaluated; five cultivars are currently grown in Greece whereas
the other two grown in India. Cultivars will be referenced to herein
as GR-1, GR-2, GR-3, GR-4, GR-5, IND-1 and IND-2. The cultivars have
early to medium maturity, and high seedcotton yielding potential
combined with good fiber quality traits. Their performance according
to each variety’s label is summarized in Table 1.
In 2003, cultivar GR-1, which is one of the most
popular cultivars among Greek farmers, was crossed as common parent
to each of the other six cultivars resulting in the following F1
combinations: GR-1 x GR-2, GR-1 x GR-3, GR-1 x GR-4, GR-1 x GR-5,
GR-1 x IND-1 and GR-1 x IND-2. During 2004 growing season,
each of the six F1 was advanced to the corresponding F2
and crossed on to its respective parents to produce the appropriate
backcross (BC1). Thus from each initial F1
cross, seed for six generations, namely P1, P2,
F1, F2, BC1P1 and BC1P2
were available. The six generations for each initial F1
cross were field evaluated during 2005 growing season in Almyros,
located in Thessaly which is the main cotton producing area in
Greece. A randomized complete block design (RCB) field arrangement
with three replications for each of the six crosses was employed.
Three row plots were used. Rows were 0.95 m spaced apart 5.0 m long.
Planting density was 15 plants m-2. Standard cultivar
practices applied to official variety evaluation trials were followed
throughout the growing season. Number of days to first open boll in
50% of the plants in each plot (DBO) was recorded as an estimate of
earliness. Individual plants were hand harvested and the sum of their
seedcotton yield was recorded on a per plot basis as an estimate of
each generation’s yielding potential. Furthermore from each
plot, a 50 boll sample was taken to determine fiber length, strength,
uniformity and micronaire using High Volume Instrument (HVI).
Conventional RCB analysis of variance was conducted separately for
each trait and for each cross. Means were separated using Fisher’s
protected LSD and/or selected one degree of freedom comparisons
(Steel and Torrie, 1980). Heterosis estimates for
seedcotton yield and for fiber quality traits were obtained as
follow: Heterosis (Het F1 and Het F2) = F1
– Mid-parent and F2 – Mid-parent,
respectively. Useful heterosis (Uhet F1 and Uhet F2)
was defined as superiority of F1 or F2 over the highest performing
parent (Meredith, 1999).
Results and discussion
The analysis of variance for seedcotton yield
produced significant generation effects in four out of the six
crosses evaluated (Table 2). Data provided evidence of effective
differentiation among some of the generations evaluated whereas
significant mid-parent F1 and F2 heterosis
effects were observed. This evidence was further confirmed from the
detailed seedcotton yield generation performance data (Table 3). The
common parent cultivar GR-1 was equivalent to GR-2, GR-4, GR-5 and
IN-1 and significantly inferior than GR-3 and IN-2 by 21.0% and 18.1%
respectively. The significant F1 and F2
mid-parent heterosis observed in the four varietal crosses ranged
from 10.0% to 24.3% and from 8.6% to 26.7% for F1 and F2
respectively. This performance was not necessarily associated with
significant difference in the yielding ability between each cross’
parents. The heterosis estimates observed were within the range of
the respective estimates summarized by Meredith (1999). On the
contrary significant useful heterosis was observed only in two
crosses. Estimates were 17.4% and 24.2% for F1 useful
heterosis in GR-1 x IN-1 and GR-1 x GR-5 crosses respectively,
whereas the corresponding F2 useful heterosis was 23.8%
and 23.2% (Table 3). Observed values were higher than those reported
(Meredith, 1999). F1 and F2 useful heterosis
are in agreement with the concept that high-yielding hybrids are
usually derived from crossing high-yielding parent (Davis, 1978).
In spite of this, no useful heterosis was observed in the crosses
between GR-1 and its respective two superior parents. These findings,
pertinent to the specific set of crosses, provide preliminary
evidence on the value of the specific cultivars to be useful
germplasm sources in applied line and/or variety development
programs. Furthermore the BC1P1 from GR-1 x
GR-3 and GR-1 x GR-5 crosses out yielding significantly the recurrent
parent GR-1 by 30.0% and 25.0% respectively whereas the BC1P2
from the GR-1 x GR-5 cross out yielded significantly the recurrent
parent GR-5 by 37.2%. This performance, although unexpected, provides
preliminary evidence on the possibility the yielding ability of the
recurrent parent to be substantially upgraded. Such data need further
testing to be confirmed before any conclusions could be reached on
their implications in applied breeding programs. Summarizing the data
previously discussed the cultivar GR-1 has good GCA and could be a
valuable germplasm source to be utilized in breeding program.
Earliness is an interesting trait, especially when
it does not sacrifice yield. Under our conditions DBO seems to be a
reliable earliness criterion. Analysis of variance produced
significant generation effects in five out of the six crosses
evaluated (Table 4). The F1 and F2 mid-parent
heterosis effect were essentially non significant. Useful heterosis
estimates for earliness, meaning earlier F1 or F2
than the common parent GR-1, were not observed. Detailed DBO
performance data (Table 5) showed that five of the cultivars were
significantly later than the common parent. The F2, BCP1
and BCP2 performance indicated that in using this material
as segregating populations, genotypes that are slightly later than
the early common parents and earlier than the late parents could be
expected.
Analysis of variance data for length, strength,
micronaire and uniformity are presented in Table 6. Significant
generation effects associated with F1 mid-parent heterosis were
observed in four out the six crosses for fiber length; for three, two
and one cross for micronaire, strength and uniformity, respectively.
Fiber data (Table 7) indicated that for the common parent cultivar
GR-1, fiber length was higher than GR-2, GR-3, GR-4, GR-5, and IND-1
and micronaire was higher than GR-4 and IND-2 but lower than GR-3. On
the contrary, no differentiation was observed for fiber strength and
uniformity with the exception of GR-1 which was higher than GR-3.
Significant mid-parent F1 heterosis was observed in four,
two, and one out of the six crosses for fiber length, both strength
and micronaire, and uniformity, respectively. The estimates obtained
were minimal and in agreement with reported data (Meredith, 1998).
The mean performance of the F2 and BC generations were not particularly encouraging in selecting genotypes with better quality
for fiber traits than parental cultivars although selection within
some of the segregating populations, which is in progress, does not
rule out identification of desirable genotypes.
Developing cotton varieties which will maintain the high yielding
ability and having the appropriate time of maturity combined with
good fiber quality traits is a breeding objective under our cotton
farming conditions. Data indicated that cultivar GR-1, used as a
common parent in this study, although seemed to be satisfactorily
meeting this requirement, could be improved in both seedcotton yield
and micronaire by using parent cultivar GR-3.
Summarizing the data discussed, preliminary
evidence indicate that some of the segregating populations derived
from the particular crosses evaluated such as the BCP1
in cross GR1 x GR3 and BCP1, BCP2 in cross GR1
x GR5 could be useful as germplasm sources in applied variety
development programs. Our research in progress, including detailed
generation means analysis and selection within the segregations
populations, might provide useful information on genetic estimates
(variance, heritability, etc) and the possibilities to effective
selection for each of the individual trait studied and / or
combinations of yielding ability with fiber quality traits and
earliness.
References
Calhoun, D.S. and D.T. Bowman. 1999. Techniques
for Development of New Cultivars. p. 361-414. In: Wayne C.
Smith(ed) Cotton: Origin, History, and Production. John Wiley &
Sons, Inc. Chaudhry,
N.R. 1997a. Cotton yields stagnating. The ICAC Recorder XV(1):3-7.
Chaudhry,
N.R. 1997b. Commercial cotton hybrids. The ICAC Recorder XV(2):3-4.
Davis, D. D. 1978. Hybrid cotton: Specific problems and potentials.
Adv. Agron. 30: 129-157.
Meredith, W. R., Jr. 1993. Registration of
‘MD51ne’ cotton. Crop Sci. 33:1415.
Meredith, W. R., Jr. 1999. Cotton and Heterosis. P. 451-462. In the
genetics and Exploitation of Heterosis in Crops. ASA-CSSA-SSSA, 677
South Segoe Road, Madison, WI 53711,USA.
Poehlman, J.M. & D.A. Sleper, 1995. Breeding Field Crops. Fourth
edition, pp. 159-172.
Steel,
R.G.D. & H. Torrie, 1980. Principles and procedures of
statistics. pp. 215-218.
Worley, S., Jr., H. H. Ramey, Jr., D.C. Harrell, and T.W. Culp. 1976.
Ontogenetic model of cotton yield. Crop Sci. 16:30-34.
Table 1. Cultivar expected agronomic performance.
Varieties Traits GR1 GR2 GR3 GR4 GR5 In1 In2 Yielding Potential H H H H H H H Earlinessz VE ME ME E ME E E Ginning % 37 38 38 36 38 35 37 Span Length (2.5%) 30.0 28.9 28.5 28.5 29.9 28.5 29.0 Strength (gr/tex)y 29 30 27 30 30 31 29 Micronaire 4.1 4.1 4.2 4.0 4.0 4.1 3.9
z VE= Very
Early, ME= Medium Early, E= Early, H= High. y
Metric mN/tex = g/tex X 9.81
Table 2. Seedcotton yield variance effects of four generations derived from six cotton cultivar crosses. Seedcotton (Mgha-1) Crosess Effects df GR1 x GR2 GR1 x GR3 GR1 x GR4 GR1 x IN1 GR1 x GR5 GR1 x IND2 Generationsz 5 ns * ns + ** ** MP vs F1 1 - + - * ** * MP vs F2 1 - * - * ** + CV(%) 11.4 6.2 6.1 11.0 4.8 5.6 z MP = Midparent y '+, *, ** effect significant at p=0.10, 0.05 and 0.01respectively.
Table 3. Seedcotton yield performance (Mg/ha) for each of four generations derived from six cotton variety along with heterosis estimates. Crosses Generationsz GR1 xGR2 GR1 x GR3 GR1 x GR4 GR1 x IN1 GR1x GR5 GR1xIND2 P1 4.8 a 4.3 c 5.2 a 4.0 bc 4.9 c 4.7 c P2 5.1 a 5.3 ab 5.0 a 3.8 c 4.9 c 5.5 ab F1 4.9 a 5.3 ab 5.2 a 4.7 ab 6.1 b 5.6 a F2 5.8 a 5.8 a 5.3 a 4.9 a 6.1 b 5.5 ab BCP1 5.1 a 5,7 a 5.0 a 4.0 bc 6.2 b 4.9 c BCP2 5.6 a 5.5 a 5.7 a 4.4 c 6.6 a 5.8 a HETEROSIS(%) F1 vs MP -1.8 ns 10.1 + 2.8 ns 20.8 * 24.8 ** 10.6 * F2 vs MP 16.4 ns 20.0 ** 3.8 ns 26.8 * 23.6 ** 8.6 + F1 vs BP -4.4 ns 0.3 ns 1.6 ns 17.4 + 25.9 * 2.3 ns F2 vs BP 13.4 ns 9.4 ns 2.6 ns 23.1 * 24.8 * 0.6 ns z MP = Mid parent, BP = Best parent. y Means within a column followed by the same letter are not significantly according to the Fisher's Protected LSD test. x '+, *, ** effect significant at p=0.10, 0.05 and 0.01respectively, one degree of freedom comparisons.
Table 4. Earliness variance effects of four generations derived from six cotton cultivar crosses. Days to boll opening from planting (DBO) Crosses Effects df GR1 x GR2 GR1 x GR3 GR1 x GR4 GR1 x IN1 GR1 x GR5 GR1 x IND2 Generationsz 5 ** ** ns * ** ** MP vs F1 1 + ns ns ns ns ns MP vs F2 1 ns ns ns ns * ns CV(%) 2.2 2.1 2.7 3.8 1.6 2.8 z MP = Midparent y '+, *, ** effect significant at p=0.10, 0.05 and 0.01respectively.
See more of Plant Breeding and Genetics
See more of The World Cotton Research Conference-4 (September 10-14, 2007)