1
if!*&~-krn9
8;: *Ma 7mm
PURDUE UNIVERSITY
GRADUATE SCHOOL
Thesis Ac,ceptance
___I \\-
h”
I’
i _,, *._.
This is to certify that the thesis irepared
l3Y
N'diapa Cisse
Enritled
. . .
.
Herltabxlrty estuuates, genetic correlation, and identification of
RAI?D markers linked to seedliug vigor and associated agronomie
traits in sorghum
Complies with University regularions and meets rhe .sTandards of the Graduare School for Dtiginaliry
and quality
*
FOC the degree of
Doctor of Philosophy
, chair
El is
?
Thi:; thesis
Irp
is net fo be regarded as confid+al.
IA-fW
.
‘----wr
1
Format Approved by:
Thesis Formar Adviser

---

HERITABILITY ESTIMATES, GENETIC CORRELATION,
AND IDENTIFICATION OF RAPD MARKE:RS LINKED TO
SEEDLIIIJG VIGOR AND ASSOCIATED AGRONOMIC TRAITS IN SORGHUM
A Thesis
Submitted to the faculty
of
Purdue University
by
Ndiaga Cisse
In Partial fulfïllment of the
Requirement for the Degree
of
Doctor of Philosophy
December 1995

---

ii
f .
I :
This dissertation is dedicated to my family,
for their support and encouragement and to ;a11 those who have made it possible.

---

.
111
ACKNOWLEDGMENTS
1 would like to express my sincere gratitude to my major professor, Dr. Gebisa
Ejeta, for his support and guidance. My thanks and appreciation are also extended to the
members of my graduate committee, Drs. Wyman Nyquist, Larry Butler and Peter
Goldsbrough for their invaluable comments and suggestions.
1 could not have proceeded with the mapping project without Mitch Tuinstra’s
guidance and advice, and 1 really appreciated his help. 1 am also grateful to the members
of the sorghum project for their assistance.
My sincere appreciation are extended to Dr. Anthony E. Hall at the University of
California at Riverside. My graduate fellowship was supported by the Bean / Cowpea
CRSP, USAID Grant No. DAN-1310-G-SS-6008-00.
The dissertation research was
supported by the International Sorghum and Millet (INTSORMIL) CRSP, IJSAID Grant
No DAN 1254-G-00-002 l-00.

---

iv
TABLE OF CONTENTS
Page
LIST O F TABLES ............................................................................................................
vii
LIST O F FIGURES.. ............................................................................................................
x
ABSTRACT.......................................................................................................................
xi
1 NTRODUCTION ...............................................................................................................
.1
LITERATURE REVIEW.. ...................................................................................................
4
Seedling vigor.. ..................................................................................
.....................
4
Seedling vigor and correlated traits......................................................................
.13
Agronomie traits.......................................................................................
.13
Phenolic compounds.................................................................................
.17
Quantitative Trait Loti.. ........................................................................................
20
CHAPTER ONE - ESTIMATES OF SEEDLING VIGOR AND THEIR GENETIC
REL<ATIONSHIP IN A RECOMBINANT INBRED LINE POPULATION OF
SORGHUM ...............................................................
,,..........................................
.30
ABSTRACT .....................................................................................
......................
.3 1
INTRODUCTION......................................................
...........................................
.33

---

V
Page
MATERIALS A N D METHODS ..........................................................................
.3 8
Plant materials.. .........................................................................................
.3 8
Field experiments.. .....................................................................................
38
Greenhouse experiment. .................................
.........................................
..3 9
Laboratory experiments.. ................................
..........................................
.40
Data analysis.. ............... .,...........................................................................
.40
RESULTS A N D DISCUSSION.. .........................................................................
.43
CHAPTER TWO - RELATIONSHIP OF SEEDLING VIGOR ESTIMATES WITH
SELECTED AGRONOMIC TRAITS I N SORGHUM.. ..................................................
.02
ABSTRACT.. .............................
...........................................................................
.63
INTRODUCTION.. .................................................................................................
6 4
MATERIALS A N D METHODS.. .......................................................................
..6 8
RESULTS A N D DISCUSSION.. .........................................................................
.7 1
CHAPTER THREE - EFFECT OF PHENOLIC CONCENTRATIONS IN SORGHUM
KERNELS ON SEEDLING VIGOR AND PLANT PRODUCTIVITY....... .<, ................. .82
ABSTRACT.. ........................................................................................................
.83
I N T R O D U C T I O N .................................................................................................
.84
MATERIALS A N D METHODS.. ........................................................
................ ..8 8
RESULTS A N D DISCUSSION.. ........................................................................
.9 1

---

vi
Page
CHAPTER FOUR - IDENTIFICATION OF QUANTITATIVE TRAIT LOCI
ASSOCIATED WITH SEEDLING VIGOR AND CORRELATED CHARACTERS
IN A RECOMBINANT INBRED LINE POPULATION OF SORGHUM.. ................. .104
ABSTRACT .........................................................................................................
10s
INTRODUCTION......................................................
..........................................
.107
MATERIALS A N D METHODS.........................................................................
1 10
D N A Preparation ....................................................................................
.1 1 0
R A P D Reactions.. ........ . .
.................................
. .
........................................
.1 1 1
Data Analysis. .............. .............................................................................
1 1 3
RESULTS AND DISCUSSION .........................................................................
.113
Seedling vigor traits.. ..............................................................................
.1 13
Adult plant traits.....................................................................................
.117
REFERENCES.. .............................................................................................................
.136
VIT,4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
...... 154

---

vii
LIST OF TABLES
Table
Page
1.1
Expected mean squares, and analysis of vari.ance for seedling vigor on visual scores
combined across years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..“........ . . . . . . . . . . . . 49
1.2
Performance of parental and recombinant inbred lines for seedling
vigor traits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
1.3
Chi-square (x’) analyses of seedling vigor scores for two genetic models...............51
1.4
Analyses of variantes of seedling vigor traits measured in the greenhouse.............52
1 .S
Analyses of variantes of IOO-seed weight and percent germination at 12” C
and 22°C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .._.................................
~ . . . . . . . 54
1.6 Genetic, additive and environmental variante estimates of seedling vigor traits.....55
1.7 Broad-sense heritability estimates of seedling vigor traits . . . . . . . . . . . . .,....“.. . . . . . . . . . . . . . . . . . . . 56
1.8
Genetic correlation coefficients and their standards errors (in parenthesis) in
S R N 3 9 x SQR recombinant inbred population . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
1.9 Genetic correlation coefficients and theirs standard errors (in parenthesis) of
seedling vigor traits measured in the greenhouse and incubator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
1.10 Genetic correlation coefficients and standard errors (in parenthesis) of seedling
visual scores traits. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . * . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...” . . . . . . . . . . . . . . . . . . . . 59
2.1
Expected mean squares, and analyses of variantes for seedling vigor visual scores,

---

ix
Page
3.7 T test analyses and standard en-ors (in parenthesis) for coleoptile color differences
in seedling vigor traits. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..~......................................I...
1 0 2
3.8 T test, analyses and standard errors (in parenthesis) for coleoptile color differences
in phenolic compounds and fïeld performances. . . . . . . . ..*.................... .d..................
1 0 3
4.1
Genetic markers that significantly cosegregrate with seedlingvigor scores
in (SRN39 x SQR) recombinant inbred population.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 1
4 . 2 RAPD markers associated with germination and emergence
in ( S R N 3 9 x S Q R ) RI population . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . .._........................
122
4.3
RAPD markers significantly associated with seedling height
in ( S R N 3 9 x S Q R ) RI population . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..“.................... 123
4 . 4
RAPD markers significantly associated with seedling dry weight
in (SRN39 x SQR) RI population ,................................................“.....................,....
124
4 . 5 RAPD markers significantly associated with plant height
in (SRN39 x SQR) RI population. Parenthesis
indica.te that marker location was not determined . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..I.................
1 2 5
4.4 RAPD markers significantly associated with maturity
in ( S R N 3 9 x S Q R ) RI population . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
4 . 7
RAPD markers signifïcantly associated with grain yield
in ( S R N 3 9 x S Q R ) RI population . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
4 . 8
RAPD markers significantly associated with phenolics compounds . . . . . . . . . . . . . . . . . . . . . . 128

---

.‘?
LIST OF FIGURES
Figure
1.1
1993 Inbred lines distribution for seedling vigor scores . . .._.........__.._._........._..._.___..
..60
1.2
1994 Inbred lines distribution for seedling vigor scores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1
4.1
A genetic linkage map of sorghurn with 59 markers in 11 linkage groups............. 130
4.2
Location of QTLs for seedling vigor scores in SRN39 .x SQR
recombinant inbred population . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .._. . . . . . . . . . . . . . . . . . 13 1
4.3
Location of QTLs for germination and emergence in SRN39 x SQR
recombinant inbred population . . . . . . . . . . . . . . . . . . ..“.............................. ~ “..........................
132
4.4
Location of QTLs for seedling height, dry matter in SRN39 x SQR
recombinant inbred population . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1 3 3
4.5
Location of QTLs for maturity, height, yield in SRN39 x SQR
recombinant inbred population . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
4.5
Location of QTLs associated with phenolic compounds . . . . . . . . . . . . . . . . . . . . . . . . . . .._.........._
135

---

xi
ABSTRACT
Cisse, Ndiaga PH.D., Purdue University, December, 1995. IHeritability estimates, genetic
cor-relations, and identification of RAPD markers linked to seedling vigor and associated
agronomie traits in sorghum. Major professor: Gebisa Ejeta.
Seedling vigor in sorghum (Sorghum bicolor (L.) Moench), is important for
improving stand establishment. These studies were conducted to investigate the
her.itabjlity of the difference in seedling vigor observed between SRN39 and Shanqui red,
1.0 assess the genetic relationships of seedling vigor with fie.ld performance and phenolic
compounds concentrations, and to identify QTLs associated with these characters. One
hundred recombinant inbred lines and their parents were evaluated for seedling vigor,
trop performance and phenolic compounds concentrations. Percent seed germination at
12” C and at 22” C, seedling emergence, seedling height and shoot dry weight were
determined. under controlled environments. Significant genetic correlation between
different estimates of seedling vigor were observed. Germination at 22” C, emergence,
seedling vigor scores, and the rate of seedling dry matter accumulation were signlficantly
correlated with grain yield. High phenolic concentrations were associated with vigorous
seedlings high percent germination at 22” C, high emergence, and taller seedlings. Only
total phenols were signifïcantly associated with grain yield. Lines with a red coleoptile

---

xii
tended to be more vigorous and more productive than lines with a green coleoptile.
RAPD markers on linkage groups D and F were significantly associated with seedling
vigor scores. Germination at low and at optimum temperatures were mostly undel
different genetic control. The marker analysis showed that the visual scoring system used
was effective in integrating germination, emergence, and seedling height. It was
concluded that Shanqui red could be a valuable parent for the development of early
maturing varieties that have improved stand establishment, and adapted to environments
where low temperatures at planting in spring and early frost in the fa11 prevail. The
identification of markers associated with seedling vigor and field performance should
make breeding for the improvement of these traits more efficient. by minimizing the
xnount of genotype by environment interaction effect.

---

INTRODUCTION
Early vigor is considered an essential component of a trop plant ideotype for a11
enviromnental conditions (Ludlow et al., 1990). In arid environments, varieties with early
vigor and good seedling establishment tend to enhance transpiration at the expense of
direct soi1 evaporation, resulting in bigh level of dry matter accumulation and improved
grain yield.
In temperate environments early planting and use of minimum tillage accentuate
germination and seedling growth problems, bec,ause low soi1 temperature and high
moisture often prevail at planting time. Seedling tolerance to low temperature is enhanced
by rapid germination, high percentage germination, and vigorous seedling growth (Keim et
al.,, 1984).
The plant characteristics that are responsible for differences in early vigor among
and within plant species have not been fùlly characterized (Acevedo et al., 1991). Some
simple characteristics, such as kernel weight, percent germination, emergence and seedling
growth (shoot height, dry weight and growth rate) may be important. Differences in
percent germination and emergence cari influence plant population density. Since trop
yield is a function of density (Willey et al., 1995!), it follows that seedling vigor cari
influence trop yield through germ&tion and emergence. Seedling vigor may have a
direct effect on the ability of the plant to accumulate dry matter. The direct effects of
seedling vigor on yield greatly depends upon the trop species. With grain crops such as
sorghum which are harvested afier they have completed their life cycle (Ml reproductive
.-.- -- ---. --
--------y------

---
---::EL&-
---

---

2
maturity), it is believed that a yield response to seedling vigor traits occurs only when
plant densities are lower than the density required to maxirnize grain yield (Tekrony et al.,
1991).
Sorghum plants produce large amounts and a great diversity of phenolic
compounds (Butler, 1989). Many of these phenols determine plant color, appearance,
nutritiona quality, and host defenses. Sorghum phenolic compounds cari be divided into
five basic groups; phenolic acids, lignins, quinones, flavonoids and tannin (Butler, 1989).
It was reported that the percent germination in sorghum cultivars with high and low tannin
content was the same (Chavan et al., 1981). IIowever both root and shoot growth were
markedly suppressed in high tannin as compared to low tannin seedlings. The rates of
germination were also the same, but the subsequent rates of root and and shoot growth
were much lower in high tannin seeds. It was concluded that tannins in sorghum seeds
retard seedling growth due to inhibition of starch degrada.tion by inactivating the
hydrolytic enzymes during germination.
Until recently, polymorphisms have been detected with phenotypic assays of
genotypes. However genetic analysis based on phenotype is a fimction of the heritability
of the trait where factors such as the environment and quantitative inheritance oRen
confound the expression of a genetic trait (Dudley, 1993). Plant breeders routinely find
that genotypes which perform well in one environment are not as well suited to another
environment. DNA markers (RFLPs, RAPDs) have advantages in that they do not bave
these interactions observed in phenotype-based assays. The identification of DNA

---

3
markers associated with seedling vigor, grain yield, plant height and maturity should allow
more efflcient breeding activities for the improvement of these traits.
Recombinant inbred (RI) lines of a cross between ‘SRN39’ and ’ Shanqui red’
(SQR), obtained through the Single Seed Descent method of plant breeding, were used to
investigate the genetic basis of the differences in seedling vigor of the two parents and to
determine the inter-relationships among different estimates of seedling vigor. The results
of these studies are presented in chapter 1. In chapter II the genetic relationships of
seedling vigor traits with yield, plant height, and maturity at high plant density
recommended for commercial production are reported. In chapter III the effects of
different phenolic compounds including tannins on seedling vigor traits, yield, plant height,
maturity, and the relationship between the different phenolic compounds and their
estimates of variante components and heritability are presented. Finally, in chapter IV of
this thesis the Quantitative Trait Loti (QTLs) associated with seedling vigor, grain yield,
plant height and maturity are presented.

---

LITERATURE REVIEW
Seedling vigor
Major constraints for dryland sorghum [(Sorghum bicdur (L.) Moench]
production in the semi-arid regions are lack of sufficient water in the seed zone and
variation in soi1 temperature at planting. Stand establishment is particularly difficult in hot
dry weather. According to Radford et al. (1989) only 55 ‘% of sorghum seed planted in
the field in Australia resulted in successful plant establishment. They estimated that
sorghum growers were losing 30 % of potential yield because of inadecluate plant density.
In sub-Saharan Afi-ica and in India, the temperature of the soi1 surface in farmers fields
commonly exceeds 45°C and temperatures as high as 60°C have been recorded (Ougham
et al., 1988). However soi1 temperatures above 45°C inhibit the germination and
emergence of grain sorghum seedlings. Peacock (1982) indicated that optimum
germination and emergence of grain sorghum occurs at soi1 temperatures of 21 to 35°C
while the lethal temperature ranged from 40 to 48°C.
Grain sorghum traditionally has not been grown in cool regions as has maize ( Zen
nmys L.). Partial credit for a broader range of adaptation for maize cari be given to
breeding efforts for improved seedling cold tolerance. Similarly improved cold tolerance
in sorghum would allow expansion of this trop into higher elevation and. more temperate
latitudes. Improved cold tolerance would also benefit production where planting is usua%

---

delayed due to danger of stand reductions from late cold periods or in cool soils
associated with minimum tillage (Bacon et al., 1986).
For decades, the studies of plant cold- tolerance and cold-stress injury had two
primary goals (Guy, 1990). The first was to describe the mechanisms during a period of
cold that ïead to ce11 injury and death These mechanisms included ice formation within
the cells and in the extracellular regions of plant tissue. The second and parallel goal, was
to catalog and understand the biochemical and physiological changes oçcurring during
cold acclimation. These studies have revealed that plant and algal cells cari rapidly begin
to alter their membrane lipid composition (Lynch et al., 1984), RNA (Cattivelli et al.,
1989) and protein content (Gilmour et al., 1988) within hours of exposure to low non-
freezing temperatures. These changes suggest a molecular basis for the adjustment of
metabolisms to low temperature. Several cDN.As for genes upregulated at low
temperature have been isolated (Schaffer et al., 1988). Guy (1990) indicated that in the
characterization of the molecular-genetic basis of cold tolerance, major opportunities exist
in targeting genes that encode cold-labile enzymes and proteins, enzymes in the
biosynthesis of lipid metabolism and key regulatory enzymes of the respiratory pathway.
Breeders have exploited the genetic variability present in many crops to develop cultivars
with improved cold tolerance.
Early vigor is considered an essential component of a trop plant ideotype in a11
environmental conditions (Ludlow, 1990). Important advantages have been attributed to
early vigor. First, the soi1 surface , which is usually moist during vegetative development,
would be shaded more by a vigorous trop and this would result in less evaporation of

---

water from the soi1 surface and therefore, more water for transpiration and growth
(Condon et al., 1987; Cooper et al., 1987; Gregory et al., 1992). Second, a greater leaf
area and growth when vapor pressure deficit is low result in greater carbon assimilation
per unit transpirational water loss than if growth occurred later when temperatures are
higher (Tanner and Sinclair, 1983). Third, a greater trop biomass by anthesis results in a
higher yield potential. In addition to increased production, more vigorous crops inhibits
the growth of weeds resulting in the use of less chemical herbicides, and may also reduce
the emerging problem of herbicide resistance in weed species (Lopez-Castaneda et al.,
1995). These reasons have led to the investigation of the factors which are responsible for
variation in early vigor (leaf area development and biomass accumulation) within trop
plants to identify ways to improve their vigor and hence water-use efficiency and yield. A
small improvement in growth rate early in the trop development cari lead to a considerable
increase in biomass at anthesis, as growth follows an exponential pattern (Richards, 1987
).
The plant characteristics that are responsible for differences in early vigor among
and within plant species have not been fùlly characterized (Acevedo et al., 1991; Regan et
al., 1992). Some simple characteristics may be important. Selecting sorghum cultivars for
rapid and uniform germination under a wide range of temperatures is important for early
seedling establishment in the field (Brar, 1994). Radford and Henzell(l990)
recommended screening commercial cultivars of sorghum for germination and seedling
elongation at temperatures occurring in seed beds at planting. At high temperature, rapid
germination cari establish the trop before soi1 drying. Genetic variability has been shown

---

7
to exist for high temperature tolerance for germination and emergence (Wilson et al., 1982
). Rapid germination permits the seminal root system to access wet soi1 ahead of the
drying front (McCovan et al., 1985 ). Using protrusion ofthe radicle as a measure of
germination, Thomas and Miller (1979) and Mann et al. (1985) reported variation in the
response of germination to temperature among diverse lines of sorghum. Lines considered
higher yielding in more tropical environments had lower base temperatures for
germination, while other lines considered higher yielding in more temperate environments
had higher base temperature. Mann et al. (1985) also reported that the response to
temperature is influenced by the environment experienced during seed maturation. More
recent studies by Lawlor et al. (1990), however, have demonstrated that the attributes
chosen as a measure of germination may influence the magnitude of any response to seed
maturation environment. When germination was defined to include protrusion of the
coleoptile, rather than the earlier protrusion of the radicle alone, no effect of the seed
maturation environment was found. Then lines no longer differed in base temperature, but
dilfered in the responsiveness of germination rate to temperature. Those genotypes
considered better adapted to more tropical environments had greater responsiveness to
tetnperature. Base temperature for germination, when defined as coleoptiie protrusion,
was higher than for radicle protrusion, and corresponded with values of about 10°C in
sorghum (Monk, 1977; Angus et al., 198 1). Defining germination to include protrusion
of the coleoptile is therefore more appropiate to field emergence, and consequently to
trop adaptation (Wade et al., 1993). Monteith (1987) proposed a linear mode1 for
considering the response of germination to temperature. In this model, germination did

---

8
not occur below a base temperature, and the rate increased linearly with temperature to an
optimum, above which it decreased linearly again to a maximum temperature above which
seeds would not germinate.
Partitioning variante components showed both base temperature for germination
defined as protrusion of the coleoptile and responsiveness of germination rate of sorghum
hybrid were subject to strong genetic control (Wade et al, 1993). Narrow-sense
heritability was low and specific combining ability high, indicating that response of
germination to temperature is hybrid specific but not constrained to hybrid groups.
Sc.reening sorghum hybrids for variation in base temperature or temperature responsivness
of germination rate was thus considered unlikely to provide a valid basis for identifying
tropically adapted germplasms. However these attributes may have value in assisting the
identification of germplasms which are desirable for trop establishment . Cold tolerance is
enhanced by rapid germination and high percentage germination. The improvement of
germination under cold temperatures was attempted in maize by combining tests in both
controlled and field environment ( McConnell et al., 1979 ). Alter four cycles of selection,
cold germination at 7.2”C was improved by about 9 % but little improvement was realized
in field emergence and seedling vigor. In an inheritance study of cold tolerance in parent
and selected populations of a Nebraska derivative of the Iowa Stiff Stalk Synthetic, Keim
et al. (1984) found that estimates of additive genetic variante for total number of plants
emerged were not significative in the parent population. Negative additive genetic
variantes were obtained for seedling vigor (visual scores) and weight per seedling.
Variante component estimates cari be negative due to sampling errors when the true

---

components are zero or small positive numbers (Nyquist, 199 1). Dudley and Mol1 (1969)
argued that repeated experimentation, involving the same ,trait in related populations, Will
give estimates which approach a “truc” value, when averaged together. If negative
estimates are not reported, an unbiased average cannot be computed from accumulated
results in the literature (Nyquist, 1991). In the selected population, a11 estimates of
additive genetic variante were positive with smaller standard errors (Keim et al., 1984).
Increased additive genetic variante may have resulted from the four cycles of selection.
Alleles for cold tolerance may have been at extremely low frequencies in the parent
population. In such a case the parent population would contain few individuals exhibiting
appreciable cold tolerance. With selection, the frequency of alleles for cold tolerance
would increase toward the value of 0.707, where additive genetic variante would be at a
maximum for complete dominante (Palconer, 1989). Additive genetic correlations in the
selected population between total number of seedlings emerged and seedling vigor, and
between emergence and weight per seedling were highly significant. The cor-relation
between seedling vigor and weight per seedling was not significant. Significant additive
genetic correlations are not surprising, since the traits a11 measured an expression of cold
tolerance.
Seedling vigor was significantly correlated with cold germination in rice (Oriza
sativa L.) ( Jones et al., 1976 ). Phenotypic recurrent selection in sorghum, under early
spring planting, resulted in a 15 % increase in cold germination afier four cycles ( Bacon,
1986 ). The sorghum seedlings emerge in the same way as maize, oats (Avena sativa L.)
and rice seedlings, by elongation of the mesocotyl and coleoptile, in contrast with barley

---

10
(Hordeunz vulgure L.) and wheat (Triticum aestivum L.) in which the mesocotyl does not
elongate ( Hosikawa, 1969 ). The mesocotyl and coleoptile of sorghum are joined by a
node which appears as a slight bulge in the Sho#ot (Mandoli and Briggs, 1984). Above this
node, the primary leaf grows within the coleoptile. A positive correlation has been found
between the coleoptile length of sorghum and its emergence (Wanjari and Bhoyar, 1980).
Wilson et al. (1982) found genetic diversity in seedling length and its response to
temperature. They also mentioned the coleoptile of sorghum, but not the mesocotyl, in
discussing the effects of temperature on emergence. However, the coleoptile generally
constitutes only a small portion of the total length of a sorghum seedling, but the
proportion increased at high and low temperatures (Radford et al, 1990). The percentage
of coleoptile in the seedling length varied from 6 to 53 % in this study with eight sorghum
genotypes at seven constant temperatures (15, 20, 25, 30, 35, 40, 45°C). The optimum
temperature for coleoptile elongation was 15 to 20°C. Both coleoptile and mesocotyl
elongation were sensitive to temperature, but mesocotyl elongation showed the largest
absolute and proportional changes between 15 and 45°C. The presence of a large
coleoptile tiller was also found to contribute substantially to early vigor in wheat (Liang
and Richard, 1994a). The Mexican dwarfwheat fails to germinate when sown in the field
as deep as the indigeneous cultivars in India mainly due to their shorter coleoptile
(Swaminathan et al., 1956). Significant differences in sorghum were found in coleoptile
length which ranged from 1.35 to 4.18 cm (Wanjari et al., 1980).
Dry matter production at the early and late vegetative stages and at anthesis was
significantly correlated in wheat (Whan et al., 1991). Similarly dry matter production was

---

11
also significantly correlated with grain yield. Good early vigor and high vegetative
biomass would then improve yield. The average heritabilities for the early vegetative, late
vegetative and anthesis stages were 72, 73, and 69 %, respectively. Fresh and dry weight
measurements showed positive combining ability effects in maize. Additive gene effect
were more important. Female effects were also significant (Barla-szabo et al., 1990).
Ram et al. (1991) estimated broad-sense heritability at 50.5 % for seedling dry weight in
pigeonpea (Cajanus Cajun L.) Millsp.. Significant differences in seedling dry weight were
found in a subset of 5 12 genotypes of the sorghum world collection (Maiti et al., 198 1).
Seedling dry weight was highly correlated to leaf area, leaf number and plant height.
The faster rates of leaf and tiller production make barley more vigorous in early
stages compared to wheat , triticale ( x Triticosecale) or oats (Avenu saliva L.) ( Lopez-
Castaneda et al., 1995). The inter-val between germination and the appearance of the first
two seedling leaves was found responsible for the substantially greater vigor of barley.
With this early vigor, barley establishes a larger leaf area and biomass faster. This has
been given as a possible reason for the success of barley compared with other cereals.
Barley is usually the highest yielding temperate cereal in low rainfall areas where there is a
Mediterranean-type climate (Gregory et al., 1992). The yield advantage of barley is
particularly evident under dry conditi,ons, but it may disappear when water is not limiting
(Fischer and Wood , 1979). However, the size of the first seedling leaf was more
important than the rate of leaf appereance, rate of tiller appearance, or partitioning of C
between different plant organs in contributing to vigor among a group of Australian and
Chinese wheat genotypes (Liang et al., 1994b). Anda et al., (1994) did not find significant

---

12
difference in leaf area of sorghum seedlings below 12°C. But raising the temperature
above 12°C changed the leaf area significantly. With an increase from 13 to 16°C in soi1
temperature, seedling leaf area increased by 12.5 % at low soi1 water and increased by
55.5 % at high soi1 water treatment. Increased leaf area was correlated with increased
shoot dry matter.
Relative growth rate is defined as the rate of dry mass accumulation per unit of
existing dry mass. Lopez-Castaneda et al. (1994) did not find relative growth rate
important in explaining the difference in seedling vigor between barley, oats, and triticale.
They concluded that difference in mass established early were maintained until
physiological maturity.
Seedling vigor is assessed using different estimates and screening techniques.
Maiti et al. (198 1) stated that seedling vigor is best assessed by direct measurement of dry
weight. However a major limitation in the assessment of dry matter production is the
difficulty in detecting clear differences due to sampling errors and the amount of work
involved in taking adequate samples. Clearly, this is a limitation if selection were to be
carried out on a large breeding populations. The use of photographs to measure ground
caver as a possible indicator of dry matter production has been attempted. This technique
was sensitive to differences in growth habit between wheat genotypes (Whan et al., 1991).
Maiti et al. (198 1) used a visual scoring system to evaluate a set of sorghum genotypes for
seedling vigor on a scale of 1 (more vigorous) to 5 (least vigorous) at 7 and 14 days afler
emergence. The system was a relative one, based on the range of variability for seedling
size in the material being used. The scoring system took into account the height of plant,

---

13
spread of leaf canopy and/or the length and width of the individual leaves. Highly
significant correlations were found between visual scoring and dry weight and leaf area.
Laboratory experiments have proven to be adequate in estimating germination.
Abdullahi et al. (1972) and Brar et al. (1994) reported a significant positive correlation
between germination in the laboratory and field emergence. Also Mendoza-Onofre et a1
(1979) reported a good correlation between performance in growth chamber and field
conditions with S 1 progenies of two sorghum populations. However McConnell et al.
(1979) attributed the lack of correlation between laboratory and field results for cold
emergence and seedling vigor to the mild spring weather during the two years of
evaluation.
Seedling vigor and correlated traits.
Agronomie traits
Seedling vigor cari be measured by many variables including kernel weight,
germination, emergence and seedling growth (shoot height, dry weight and growth rate).
Many studies emphasized the relationship of laboratory germination and vigor to field
emergence. There is less published information relating seedling vigor to other aspects of
trop performance (Tekrony et al., 199 1). However there are possibilities that seedling
vigor cari influence trop yield (Ellis, 1992). Differences in percent germination and
emergence cari influence plant population density. Since trop yield is a fùnction of density
(Willey et al., 1969), it follows that seedling vigor cari influence trop yield through
germination and emergence. If seed quality (size, percent germination) only affected

---

1 4
percent emergence, then growers could theoretically, overcome such effects by adjusting
seed sowing rate. It has been shown that the effect of differences in laboratory
germination on field emergence in different seedbeds cari be quantified (Khah et al., 1986;
Wheeler et al., 1992). However, in practice, adjustments in seed sowing rates are
hampered by diffrculties in forecasting the particular seedbed environment (Ellis, 1992).
Seedling vigor also affects trop perlormance through effects on the plant growth
processes involved in the production of yield. Yield of any trop is determined by the solar
radiation intercepted by the plant community, the effrciency with which intercepted
insolation is converted to dry matter, and the proportion of the biomass that is economic
yield (Charles-Edwards, 1982). Most of the pla@ tissues involved in the production of
dry matter and yield are formed after seedling emergence, and it seems unlikely that seed
qu.ality (weight, percent germination) would influence their ability to carry out
physiological processes and accumul,ate dry matter. Seed quality did not affect the relative
growth ra.te of soybean seedlings, provided they were free of physiological injury or
necrotic lesions (Egli et al., 1990). Priming or natural variation in seed quality have been
reported to have no effect on the relative growth rate of orrions (Allium cepa L.; Ellis,
1989). However genetic aberrations, which may occur in long term storage, could cause
impaired physiological function in later formed plant tissue (Roberts, 1972; Harrison,
1966).
Seedling vigor may have a direct effect on the ability of the plant to accumulate
dry matter. The direct effects of seedling vigor on yield greatly depends upon the trop
species. Crop harvested during vegetative growth are frequently planted at low

---

population densities and harvested on an individual plant or area basis as vegetative mass
of aboveground [Lettuce (Lactuca diva L.), cabbage (Brassica oleracea var capitata)]
or underground [sugarbeet (Beta vulgaris L.), radish (Rajbms sativua L.), carrot (Dam~s
carota L.)] structures. The effects of seedling vigor cari be specially critical in these
crops, where delayed emergence or missing plants may reduce yield and uniformity at
harvest (Tekrony et al., 1991).
Crops harvested at an early stage of reproductive development usually are planted
at higher population densities than crops harvested during vegetative growth. The
quantity (fresh weight), uniformity, and quality of reproductive structures determine the
yield. Perry (1969) evaluated high and low-vigor seed lots of two pea cultivars and
reported significantly lower yield for low-vigor seed lots tbinned to similar plant
population as high-vigor lots. Abdalla and Roberts (19691) stored pea seeds in various
artificial environments, and reported that seed lots with lower viability had lower early
plant growth rates (O-35 days afier planting) and reduced plant height and leaf number at
45 and 59 days afier planting. At later stages of growth, no significant differences in
relative growth rate were recorded, and reductions in dry seed yield occurred only for
those seedlots where viability had declined below 50%. Unfortunately, measurements of
fresh seed and fruit weight at an early stage of reproductive development were not taken.
A non significant relationship between germination and fresh pod weight was reported in
lima beans (Phaseohs Zinerzsis L.; Bennett and Waters, 1984).
Grain crops are harvested alter they have completed their life cycle (full
reproductive maturity) and only the seeds are harvested for yield. Work with soybean

---

16
(Tekrony et al., 1987) bean (Spilde, 1987) and corn (Abegbuyi et al., 1989) suggest that
there is no relationship between germination, seedling dry weight and seedling growth rate
with yield. However Sign&ant increases in yield with seedling growth rate were shown
for corn (Burris, 1975). Significant increases in yield with high percent, germination seed
were also shown for corn when grown at low population densities compared with seed
that had been stored for 5 to 7 years (Funk et al., 1962). It was demonstrated that
seedling growth rate of spring barley showed an advantage only at lower plant density,
while no association existed at normal population (Perry, l980). It was found that low
percent-germination seeds of spring wheat produced lower yields only in lower than
normal populations or later than normal plantings. Yield of soybean was reported to be
related to accelerated aging, cold test, seedling and seedling growth rate in hi11 plots but
not in row plots planted at normal populations. In sorghum a significant relationship
between yield and speed of germination, and root and shoot growth was reported
(Camargo et al., 1973) but, the plant density (66,000 plants / ha) was lower than normally
recommended (Vanderlip, 1972). No significant differences were found between kernel
weight with days to 50 % bloom, plant height, and grain yield in sorghum (Suh et al.,
1974). It was also found that kernel weight has no influence on grain yield of sorghum
(Maranville et al., 1977).
Phenolic compounds
Sorghum plants produce large amounts and a great diversity of phenolic
compounds (Butler, 1989). Many of these phenols determine plant color, appearance,

---

17
nutritional quality, and host defenses. Polyphenols are secondary metabolites. Their
amount and nature vary greatly with genotypes and environmental conditions under which
plants are grown.
Chemically, phenolic acids are the simplest polyphenols of sorghum. Hahn et al.,
(1983:) identified eight different phenolic acids in grain sorghum. Furrilic acid was the
most abundant of these compounds. In addition, 12 other unidentified peaks were
separated by high-performance liquid chromatography.
Anthocyanidins are the major pigments in most plants. In sorghum the dominant
pigments are the 3-deoxyanthocyanidins. The color of sorghum kernels is influenced by
pericarp color, mesocarp thickness, presence of testa, and by endosperm texture and color
(Hahn and Rooney, 1986). The pericarp color is determined by two genes, (Kambal et al.,
1976) and cari be white, lemon-yellow or red. Kambal et al. (1976) did not find any
visible pigments in white-seeded kernels but considerable amounts ofp-coumaric, caffeic
and ferrulic acids were detected. The pigment in the yellow kernels was identified as
eriodictyol chalcone, a deep yellow pigment. The red seeds contained the anthocyanidins,
luteolinidin and apigeninidin. Doggett (1988) classified sorghum seedlings in two groups,
red and green. Red coleoptile color is controlled by a single dominant gene over green.
Flavan-4-ols, also called leucoanthocyanidins since they are converted to
anthocyanidins when heated in acid with the loss of a water molecule (W’atterson and
Butler, 1983) include monomers of flavanols such as flavan-3,4 diols and flavan-4-01s.
The concentration of flavan-4-01s in sorghum seeds is highly dependent on the seed
maturity (Jambunathan et al., 1990). Grain at early stages of maturity (10 and 14 days

---

18
afier flowering) contained the highest flavan-4-01 concentrations, followed by a drastic
decrease with increased maturity. It was suggested that flavan-4-01s could be degraded,
converted or incorporated into other molecules such as 3-deoxyanthocyanidins or tannins.
Tannins are a group of phenolic compounds found in plants, which convert animals
hides to leather during the tanning process (Butler, 1989). There are two classes of
tannins. Hydrolysable and condensed tannins. Only condensed tannins, which are
oligomers of flavan-3-01s have been found in sorghum. These oligomers are now referred
to as procyanidins, because the red anthocyanidin pigment cyanidin is released when
treated with minera1 acids. Tannins are the most abundant phenolic coumpound
extractable from the seed of brown, bird-resistant sorghum (Hahn et al., 1984). Tannins
bind to and precipitate proteins, reducing the nutritional value of the grain. High tannin
sorghums have different kernel structures from other sorghums (Hahn et al., 1984). High
tannin sorghums have a prominent pigmented testa located beneath the pericarp. The
pigmented testa is purple or reddish-brown and varies in thickness. The presence of a
pigmented testa is controlled by the complementary Bl and B2 genes. The S gene
controls the presence of pigments and tannins in the epicarp. When S is dominant, more
phenols and tannins are in the pericarp.
It was reported that the percent germination in sorghum cultivais with high (3.4
%) and low (0.5 %) tannin content was the same. However both root and shoot growth
were markedly suppressed in high tannin as compared to low tannin seedlings. The rates
of germination were also the same, but the subsequent rates of root and and shoot growth
were much lower in high tannin seeds (Chavan et al., 1981). The tannin content decreased

---

19
markedly during germination. Tannins are located in the seed coat of the sorghum grain
(Jumbunathan et al., 1973). The ioss of tannin was attributed to leaching in growth
medium and penetration into the endosperm with imbibed water during germination.
Starch content decreased, and the rate of formation and total accumulation of reducing
sugars and free amino acids were lower in high tannin seeds. The interpretation was that
starch and protein degradation were inhibited or lowered in high tannin seeds during
germination, leading to suppressed seedling growth. This inhibition would result from the
portion that enters the endosperm. Such tannins are likely to form complexes with seed
protein reserve and enzymes, and inactivate them (Chavan et al., 1981).
During germination, reserves of nutrients like star-ch and proteins are degraded to
soluble sugars and amino acids, respectively, to meet the seedling requirements. Any
depression of starch and protein degradations would indicate interference with the
metabolic systems operating on reserve starch and protein, mainly enzymes like amylases
and proteases (Dalvi, 1974). Tannins are reported to for-m complexes with hydrolytic
enzymes and inactivate them (Tamir et al., 1969; Milic et al., 1972). A marked
suppression of seedling root growth was also observed with a low tannin (0.1 ‘A) sorghum
cultivar, germinated at 1 %, 2 % and 3 % tannin acid concentrations. The inhibition
increased with concentration and time (Chukwura et al., 1982). A decrease in starch
content in the control sample (distilled water) and 1 % tannic acid solution, but not at
higher concentrations was also noted. A concommitant increase of soluble carbohydrate
content at low concentration of tannic acid, and in distilled water , and a decrease at
higher concentrations were also observed. The fact that high concentrations of tannic acid

---

reduce the level of soluble carbohydrate falls germination, below its original level, was
viewed as an indication that tannic acids directly inhibited the production of these
carbohydrates. Alpha and beta-amylase activity were also observed to be inhibited by an
increase in the concentration of tannic acid. It was concluded that tannins present in
sorghum seeds retard seedling growth due to inhibition of starch degradation by
inactivating the hydrolytic enzymes during germination.
Quantitative Trait Loti
Improvement of trop species for quantitative traits is difficult because the effects
of individual genes controlling the traits cannot be readily identified. Mather and Jinks
(197 1) summarized several cases where simply inherited markers were associated with
variation in quantitative traits, Sax (1923) first reported association of a simply inherited
genetic marker with a quantitative trait in plants when he observed segregation of seed
size associated with segregation for a seed coat color marker in beans ( Phaseolus
vulgaris L.). Rasmusson (1935) demonstrated linkage of ,flowering time ( a quantitative
trait ) in pea ( Pisum sativunz L. ) with a simply inherited gene for flower color. Everson
and Schaller ( 1955 ) found morphological markers which flanked a chromosomal region
atfecting yield in barley. In maize, translocations have been used to assoçiate segregation
for quantitative traits such as European cor-n borer resistance with individual chromosome
segments ( Burnham, 1966 ). In wheat, monosomics have been used to identify
associations of quantitative traits with individual chromosomes ( Law, 1967 ).

---

21
Though these markers have served well in various types of basic and applied
research, their use in many areas of plant breeding has been very limited. Major
limitations of these studies included the limited number of markers available, undesirable
effects of many of the morphological markers, and in the case of translocation or whole
chromosome effects, the extreme size of the chromosome being compared.
The development in recent years of protein and DIVA markers offers the possibility
of developing new approaches to breeding procedures. The greater utility of molecular
markers derives from five inherent properties that distinguish them from morphological
markers ( Stuber, 1992 ): (1) Genotypes of molecular loti cari be determined at the whole
plant tissue and cellular levels. Phenotypes of most morphological markers cari only be
distinguished at the whole plant level, and frequently, the mature plant is required. (2) A
relatively large number of naturally occuring alleles cari be found at molecular loti.
Distinguishable alleles at morphological marker loti occur less frequently and oRen must
be induced through the application of exogenous mutagens or the construction of special
genetic stocks. (3) Deleterious effects are not .Usually associated with alternate alleles of
molecular markers. This is not the case with morphological markers, which are often
accompanied by undesirable phenotypic effects. (4) Alleles of some molecular markers (
RFLP, SSR ) are codominant, allowing a11 possible genotypes to be distinguished in any
segregating generations. Alleles at morphological marker loti usually interact in a
dominant-recessive manner, prohibiting their use in many crosses. (5) Unfavorable
epistatic mteractions frequently occur among loti encoding morphological marker-traits
thereby limiting the number of segregating markers that cari be tolerated or unequivocally

---

22
scored in a single population. Most molecular markers appear to be fiee of epistatic
eftects, Thus the number of loti that, cari be monitored in a single population is
theoretically unlimited. Limitations are however dictated by the number of polymorphie
markers for which assay procedures are available or by the limitations associated with the
facilities and resources available to the researcher or breeder.
Isozymes were the first molecular markers used to identi@ QTL in maize, tomato (
Lycupemicon spp. ), wild oats ( Avena fatua L. ) and soybeans [ Glycine max ( L. ) Merr.
1, ( Stuber, 1992 ). The effectiveness of marker-assisted selection in maize was
demonstrated with the use of isozyme markers ( Stuber and Edwards, 1986 ). Although
isozyme markers likely have no phenotypic effects, the numbers of such markers available
are small.
The large numbers of restriction fragment length polymorphisms (RFLP ) and
random amplified polymorphie DNA ( RAPD ) markers in many species has allowed the
development of linkage maps with a high degree of resolution. A primary genetic linkage
map, consisting of easily scored polymorphie marker loti uniformly distributed throughout
a genome, is an essential prerequisite for detailed genetic studies and marker-facilitated
breeding approaches in any trop plant ( Stuber, 1992 ).
Construction of a linkage map involves following the inheritance of the markers in
appropiate pedigrees. Either the F2 produced from crossing two lines or the backcross of
the Fl to one of the parents provides an appropiate mapping population. Although more
complex to analyse, an F2 provides almost twice as much information as a backcross
because markers are segregating in both the male and female gametic populations

---

23
generating the F2 ( Lander et al., 1987 ). Recombinant inbred lines ( RI ) also are useful
for generating linkage map (Burr et al., 1988 ).
Recombinant inbred lines are produced by continually selfing or sib-mating the
progeny of individual members of an F2 population until homozygosity is achieved.. Each
Rl line is fixed for a different combination of linked blocks of parental alleles, SO an RT
family constitutes a permanent population in which segregation is fixed. If the original
parents of the F2 were inbred lines, then only two alleles at a given locus Will be
segregating in the population. RI populations offer two major advantages over F2 or
backcross populations. First, once homozygosity has been attained, RI lines may be
propagated indefinitely without further segregation. This is beneficial because in any
major mapping effort, DNA is eventually exhausted when either an F2 or backcross
population is used. In order to resume mapping, the allelic distribution of a11 markers must
be redetermined in a new segregating population. Second, RI lines undergo multiple
rounds of meiosis before homozygosity is reached. As a result, linked genes have a
greater probability of recombination. Linkage beyond 20 CM is frequently not detected
because of the extensive recombination characteristic of RI lines ( Burr, 1991 ).
Several sorghum linkage maps have been published [ Hulbert et al. ( 1990 );
Whitkus et al., ( 1992 ); Melake Berhan et al. ( 1993 ); Ragab et al. ( 1994 ); Chittenden
et al., ( 1994 ); Pereira et al., ( 1994 ) and Weerasuriya ( 1995 ) 1. These studies have
shown that single-copy maize sequences hybridized well to sorghum DNA, and detected
low copy number bands in sorghum. However maize repetitive DNA sequences
hybridized poorly, or not at all, to sorghum. This suggests that the larger size of the maize

---

2-i
genome is due not to the number or types of genes, but rather to the differences in the
amount of repetitive DNA ( Bennetzen et al., 1993 ). The linkage relationships of
polymorphie loti in maize and sorghum were usually conserved. Probes that were linked
in maize were ofien linked in sorghum, with large regions of colinearity between the
genomes. However, several rearrangements were also detected. Whitkus et al., ( 1992 )
suggested that the primary process involved in the divergence of maize and sorghum
genomes were duplications, inversions and translocations. Chittenden et al., ( 1994 ) have
published the most “extensive” genetic map of sorghum to date. Cosegregation of 276
KFLP loti revealed a genetic map comprised of 10 linkage groups putatively
corresponding to the ten gametic chromosomes of 5. bicolor and Spropirzquum.
This
map spanned a genetic distance of 1445 CM with an average of 5.2 CM between markers
and was estimated as an 93 % soverage of the sorghum genome.
L
The underlying assumption of using marker loti to detect polygenes is that linkage
disequilibrium exists between alleles at the marker locus and alleles of the linked
polygene( s). Linkage disequilibrium cari be defined as the nonrandom association of
alleles at different loti in a population. While a recombinant inbred populations have less
linkage disequilibrium compared to backcross or F2 populations, due to more opportunity
for meiotic recombination, they have the advantage of homozygous lines that cari be
replicated and retested for more accurate measurement of the quantitative trait ( Burr,
1991).
There are several statistical procedures for determining whether a polygene is
linked to a marker gene, and they a11 share the same basic principle: TO partition the
m*
--^_----
.-.-B-I

---

25
population into different genotypic classes based on genotypes at the marker locus and
then to use cor-relative statistics to determine whether the individuals of one genotype
differ significantly compared to individuals of other genotype(s) with respect to the trait
being measured ( Tanksley, 1993 ). If the phenotypes differ significantly, it is interpreted
that a gene(s) affecting the trait is linked to the marker locus used to subdivide the
population. The procedure is then repeated for additional marker loti throughout the
genome to detect as many loti as possible. Normally, it is not possible to determine
whether the effect detected with a marker locus is due to one or more linked genes
atfecting the trait. For this reason, the term quantitative trait locus ( QTL ) was coined to
describe a region of a chromosome ( usually defined by linkage to a marker gene ) that has
a significant effect on a quantitative trait.
The simplest approach for detecting QTL is to analyse the data using one marker
at a time. This approach is often referred to as single point analysis and does not require a
complete molecular marker linkage map. The disadvantages of point analysis have been
summarized by Tanksley (1993) as: (a) The farther a QTL is from the marker gene, the
less likely it is to be detected statistically due to crossover events between the marker and
QTL that result in misclassification. (b) The magnitude of the effect of any detected QTL
will normally be underestimated, due also to recombination between the .marker locus and
QTL. Both problems are, however, minimized when a large number of segregating
molecular markers are used, covering the entire genome (usually at inter-vals less than 15
CM ). Under these conditions any potential QTL would be closely linked to at least one
molecular marker. For progenies resulting from crosses between homozygous lines, only

---

two marker genotypes are available for comparison in backcross or recombinant inbred
populations. Thus F tests in the analysis of variante or t test between marker genotyge
means are appropriate. Alternatively, linear regression of marker genotype means on
genotype cari be used to estimate the additive effect associated with the marker locus (
Dudley, 1993 ).
Distributional extremes or trait-based analysis ( Lebowitz et al., 1987 ) is a
modification of the single point approach. In this analysis, individuals in the tails of an F2
or backcross distribution are sampled for marker frequencies. Those markers differing in
frequency between the tails are assumed to be associated with a QTL affecting the trait.
The method is effective only for only one trait at a time because different individuals Will
1ik:ely be in tails for different traits. Likewise, it is likely to be effective only for traits
controlled by a small number of QTL which show little interaction.
Lander and Botstein (1989 ) proposed a method called interval analysis to take
ml1 advantage of linkage maps for quantitative studies. Instead of analysing the
population one marker at a time, sets of linked markers are analysed simultaneously with
re(gard to their effects on quantitative traits. By using linked marker analysis, it is possible
to compensate for recombination between the markers and the QTL, increasing the
probability of statistically detecting the QTL and also providing an unbiased estimate of
the QTL effect on the character. The major benefit of inter-val analysis over the point
approach is tealized when linked markers are fairly far apart ( >20 CM ). Under these
conditions there are likely to be many crossovers between the markers and QTL, which
cari be compensated for with interval analysis ( Tanksley, 1993 ). When the marker

---

27
density is higher ( markers < 15 CM apart ) point and interval analysis give nearly identical
results. When marker loti are very far apart (e. g. > 35 CM ), even inter-val analysis is
ineficient in detecting QTL in the interval between the marker loti. Because maximum
likelihood estimates reduce to least squares estimates when data are normally distributed
(Snedecor and Co&ran, 1989), Paterson et al. (1988) and Bubeck et al. (1993) showed
that both analytical methods gave virtually the same results in detecting QTL.
Haley and Knott ( 1992 ); Martinez and Curnow ( 1994 ); Wright and Movers
( 1994 ) used inter-val mapping by the regression approach and obtained similar results as
Lander and Botstein ( 1989 ) with the log-likelihood analysis. The relative simplicity and
computational rapidity of the regression method made it easier to fit models for two or
more linked and / or interacting QTL, and gave good estimates of QTL effects. However
the effrciency of flanking-marker methods decreases as the number of incompletely-
ge:notyped individuals increases. For this reason Kearsey and Hyne ( 1994 ) proposed a
simple “marker-regression”approach. This approach produced estimates of QTL locations
and effects comparable to interval mapping based on log-likelihood or multiple regression.
The method is simple to understand, easy to implement and offers unique features: First
the residual mean square cari be used to test the adequacy of the simple one-QTL mode1
on a given chromosome in a single test. Second, it provides a simple test for whether the
QTL, located on a given chromosome in different populations, are the same and this is
achieved by standard joint regression analysis. Different populations typically segregate at
different marker loti and hence create difficulties for flanking-marker methods. The
method is applicable to F2, backcrosses, doubled haploid hnes or recombinant inbred lines.

---

28
Dudley ( 1993 ) stated that with large numbers of markers and a single factor
analysis of variante for each marker locus, a certain proportion of the effects Will be
declared significant when in fact there is no association between the marker and QTL
( false positive or type 1 errors ). For example, if the 0.05 probability level is used, five
out of 100 independent F tests are expected to be Sign&ant even if none of the marker
loti is significantly associated with a QTL. Because large numbers of markers are ofien
used ( e.g > 100) and a number of traits may be analysed, an appreciable number of type 1
en-or is likely. Lander and Botstein ( 1989 ) suggest using a significant level equivalent to
0.001 in order to reduce the number of false positives. Their argument is on reducing the
probability of finding a type 1 error any place in the genome to O.OS when analysing a
species with 12 chromosomes and markers uniformly spaced 20 centiMorgans apart. But
reducing the probability of type 1 error increases the probability of type 2 errors
( rejection of the hypothesis of no association between a marker and a QTL when in fact
such an association exists ). The problem of determining appropriate threshold values has
several other sources: There is the question of determining ( or approximating ) the
distribution of the test statistic under an appropriate nul1 hypothesis. Also the sample size,
the genome size of the organism under study, the genetic map density, segregation ratio
distortions, the proportion and pattern of missing data and the number and magnitude of
segregating QTL also contribute to the problem.
Churchill and Doerge ( 1994 ) described a procedure for estimating a threshold
value and thus detecting significant QTL effects which is valid for any continuously
distributed trait. The method gives the correct type 1 error level and has power to detect

---

29
QTL effects under the alternative hypothesis. The method is based on the concept of
permutation tests ( Fisher, 1935 ). It involves repeated “shuffling” of the quantitative trait
value and the generation of a random sample of the test statistic from an appropiate nul1
distribution.
Recently Kruglyak and Lander ( 1995 ) extended interval mapping to any
quantitative trait regardless of its distribution, through the use of another nonparametric
method. The basic statistic Zw is a generalization of the Wilcoxon rank-sum statistic to
the situation of interval mapping. The effrciency of this test relative to the t-test is 96 %, if
the distribution is normal and is never less than 86 % for any other distribution
(
Lindgreen, 1968 ). The efficiency is defined as the ratio of the sample sizes required for
the two tests to achieve the same power. This approach has been incorporated in
MAPMAKER / QTL ( version 2 ), allowing robust mapping of QTL without concern
about the precise distribution of the trait.

---

30
CHAPTER ONFi
ESTIMATES OF SEEDLING VIGOR
AND THEIR GENETIC RELATIONSHIP IN A
RECOMBINANT INBRED POPULATION OF SORGHUM

---

31
ABSTRACT
Seedling vigor in sorghum (Sorghum bicolor (L.) Moench) is important for
improving stand establishment in arid regions, and where low soi1 temperatures prevail at
planting time. This study was conducted to evaluate different methods of estimating
seedling vigor, and to assess the genetic basis of the difference in seedling vigor observed
between two cultivars, SFW39 and Shanqui red. One hundred recombinant inbred lines
generated from the above parents, and five controls, were evaluated for seedling vigor in
the field for two years. A visual score of 1 (most vigorous ) to 9 (least vigorous) was
used. The experimental materials were also evaluated in an incubator for percent
germination at 12” C and at 22” C, and in the greenhouse for emergence, seedling height
and shoot dry weight. The two-year field scores were highly correlated (r = 0.64, P =
0.000 l), indicating that visual scoring was a reliable method of estimating seedling vigor.
Shanqui red had a mean score of 1.8 in 1993 and 1 in 1994, while SRN39 had respectively
as scores 7.9 and 7 for the two years. the inbred line distribution varied between 1.6 and
7.4 in 1993, and between 1 and 9 in 1994. The 1993 data fit a duplicate dominant mode1
(x2 = 0.053, 0.9 < P < 0.75), while a triplicate dominant mode1 was appropiate for 1994
(x2 = 1.12, 0.5 <P < 0.25). Thus seedling vigor in this sorghum population is controlled
by only a few genes, probably only two or three. Sign&ant genetic and additive variantes
were observed for the visual scores and the traits measured in controlled environments.
The genetic correlation coefficients of seedling scores with the different estimates of vi-or
were Sign&ant except withlOO-seed weight. Signiticant genetic interrelationships were

---

32
also revealed between the traits measured in the greenhouse and incubator. It was
concluded that kernel weight has little effect on stand establishment. The relatively simple
inheritance of the visual scores, combined with the identification of molecular markers
associated with this trait, indicates that seedling vigor cari be used efficiently by breeding
programs geared towards improving stand establishement.

---

33
INTRODUCTION
Early vigor is considered an essential component of a trop plant ideotype for a11
environmental conditions (Ludlow et al., 1990). In arid environments, varieties with early
vigor and good seedling establishment tend to enhance transpiration at the expense of
direct soi1 evaporation, resulting in high level of dry matter accumulation and improved
grain yield. In temperate environments early planting and use of minimum tillage
accentuate germination and seedling growth problems, because low soi1 temperature and
high moisture oRen prevail at planting time. Seedling tolerance to low temperature is
enhanced by rapid germination, high percentage germination, and vigorous seedling
growth (Keim et al., 1984).
The plant characteristics that are responsible for differences in early vigor among
and within plant species have not been fi~lly characterized (Acevedo et al., 1991; Regan et
al., 1992). Some simple characteristics may be important. Evans et al. (1977) reported
that wheat seedlings grown from large seeds accumulate more dry matter than seedlings
grown from small seeds. More recently, Lafond et al., (1986) reported that seed size and
speed of emergence contributed to differences among nine wheat cultivars in seedling
vigor measured as shoot dry weight. Seed size accounted for 50 % of the variation in
seedling shot dry weight. It was concluded that improvement of seedling vigor could be
done by selecting for seed size, speed of emergence and / or rate of plant development.
Ram et al., (1991) also reported that seed weight in pigeonpea was significantly correlated
with germination percentage, field emergence, seedling dry matter and two indices of

---

34
seedling vigor. Kernel weight of sorghum has also been shown to influence the
percentage of germination (Abdullahi and Vanderlip, 1972; Maranville and Clegg, 1977)
seedling weight (Swanson and Hunter, 1936) and stand stablishment (Abdullahi and
Vanderlip, 1972), while other studies have found kernel weight to be poorly related to
tield establishment (Vanderlip et al., 1973). Selecting sorghum cultivars for rapid and
uniform germination under a wide range of temperatures if; important for early seedling
establishment in the fïeld (Brar, 1994). Rapid germination permits the seminal root
system to access wet soi1 and establish the trop before soi1 drying (McCovan et al., 1985).
Radford and Henzell(l990) recommended screening commercial cultivai-s of sorghum for
germination and seedling elongation at temperatures occuring in seed beds at planting.
Genetic variability has been shown to exist for high temperature tolerance for germination
and emergence (Wilson et al., 1982). Using protrusion of the radicle as a measure of
germination, Thomas and Miller, (1979) and Mann et al. (1985) reported variation in the
response of germination to temperature among diverse lines of sorghum.
In tropical maize populations, the rate of cold emergence was positively correlated
with seedling vigor as a result of genetic recombination and selection for adaption to their
environment (Blum, 1988). The improvement of germination under cold temperatures
was attempted in maize by combining tests in both controlled and field environments
(McConnell et al., 1979). Alter four cycles of selection, cold germination at 7.2”C was
improved by about 9 % but little improvement was realized in emergence and seedling
vigor in the fïeld. In rice, seedling vigor was significantly correlated with cold germination
and was common in varieties adapted to high altitude (Jones et al., 1976). Phenotypic

---

35
recurrent selection in sorghum under early spring planting resulted in a 15 % increase in
cold germination after four cycles (Bacon, 1986). Laboratory experiments have proven to
be adequate in estimating germination, and establishing a significant positive correlation
between germination in the laboratory and field emergence (Abdullahi et al., 1972,
Mendoza-Onofie et al., 1979, and Brar et al., 1994). Sorghum seedlings emerge by
elongation of the mesocotyl and coleoptile, the same way as in maize, oats and rice, in
contrast with barley and wheat in which the mesocotyl does not elongate (Hosikawa,
1969). Stand establishement of grain sorghum is oRen a problem in crusted soils which
are common in semi-arid climates. Emergence potential in crusted soils was found to be
correlated with coleoptile diameter and potential germination (Mason et al., 1994).
There have been few reports on the genetics of seedling vigor. Fresh and dry
weight measurements of maize seedlings showed positive çombining ability effects.
Additive gene effects were important. Female effects were also significant (Barla-szabo et
al.., 1990). Ram et al., (1991) estimated broad-sense heritability at 50.5 o/O for seedling dry
weight in pigeonpea (Cajanus cajan .L.) Millsp. A positive correlation bas also been
found between the coleoptile length of sorghum and its emergence (Wanjari and Bhoyar,
1980). A narrow-sense heritability of 0.3 1 for coleoptile length was reported in wheat
(Nykaza, 1983). Significant additive genetic effects for coleoptile length in durum wheat
(Triticum durum, Desf.) was also found (Ouassou et al., 1991). Soman et al., (1991)
found significant differences among Pearl millet genotypes for coleoptile and mesocotyl
length, Genetic variation for coleoptile length was found in two Pearl millet populations
Tifi #2 S 1 (Tifi2) and Nebraska Early Dwarf Synthetic (M’Ragwa et al., 1995). There

---

were significant responses to selection in Tift2 population for coleoptile length evaluated
in germination towels. Realized heritabilities of TiR2 population for long and short
coleoptile length were 0.21 and 0.55. For the Nebraska population realized heritabilities
for long and short coleoptile were 0.00 and 0.90, respectively. Wilson et al. (1982) found
c
genetic diversity in seedling length and its response to temperature. They also mentioned
the coleoptile of sorghum but not the mesocotyl in discussing the effects of temperature
on emergence. However, the coleoptile generally constitutes only a small portion of the
total length of a sorghum seedling but the proportion increased at high and low
temperatures (Radford et al, 1990). The percentage of seedling length contributed by the
coleoptile varied from 6 to 53 % in this study with eight sorghum genotypes at seven
constant temperatures (15, 20, 25, 30, 35, 40, 45OC). The optimum temperature for
coleoptile elongation was 15 to 20°C. Both coleoptile and mesocotyl elongation were
sensitive to temperature, but mesocotyl elongation showed the largest absolute and
proportional changes between 15 and 45°C. The presence of a large coleoptile tiller was
also found to contribute substantially to early vigor in wheat (Liang and Richard, 1994a).
The Mexican dwarf wheat fails to germinate when sown in the fïeld as deep as the
indigeneous cultivars in India mainly due to their shorter coleoptile (Swaminathan et al.,
1956). Significant differences in sorghum were found in coleoptile length ranging from
1.35 to 4.18 cm (Wanjari et al., 1980).
Crop growth rate or simply growth rate (GR) is detined as the increase of plant
material per unit of time, while relative growth rate (RGR) is the rate of dry mass
accumulation per unit of existing dry mass. Growth rate (GR) cari always be applied

---

37
without becoming involved in any assumption about the for-m of the growth curves; it also
has as advantage that meaningful values cari be obtained even when it is possible to take
only two harvests (Radford, 1967). Lopez-Castaneda et al., (199417) did not find relative
growth rate important in explaining the difference in seedling vigor between barley, oats,
and Triticale. They concluded that differences in mass established early were maintained
until physiological maturity.
Seedling vigor is best assessed by direct measurement of dry weight. Seedling dry
weight was highly correlated to leaf area, leaf number and plant height in sorghum (Maiti
et a1.,1981).
However a major limitation in the assessment of dry matter production is the
diffrculty in detecting clear difference due to sampling errors and the amount of work
involved in taking adequate samples. Maiti et al., (1981) used a visual scoring system to
evaluate sorghum genotypes for seedling vigor on a scale of 1 (more vîgorous) to 5 (least
vigorous) at 7 and 14 days after emergence. Highly significant correlations were found
between the visual scores and dry weight and leaf area.
Our studies were prompted by field observations at West Lafayette, Indiana, that
Shanqui red sorghum excelled over other lines in its ability to initiate vigorous seedling
growth in the spring. The objectives of the study reported in this chapter were to
determine if the difference were genetic and to establish the interrelationships among
different estimates of seedling vigor in a recombinant inbred population derived from the
cross between ‘SRN39’ and ‘Shanqui red’.

---

38
MATERIALS AND METHODS
Plant materials
Recombinant inbred (RI) lines were obtained, through the Single Seed Descent
method of plant breeding from a single cross between SRN39, an Aftican caudatum
genotype, and Shanqui red (SQR), a Chinese Kaoling line. These two parents differ in
seedling vigor with Shanqui red being more vigorous. One hundred random F2 plants of
this cross were selfed and advanced by Single Seed Descent to the F5 generation. Seeds
from these plants were then used to grow Fd progeny rows in Puerto Rico. Several plants
within a row were selfed and bulked to produce F7 seeds that were used for a two-year
experiment in the field. Selfed F8 seeds obtained in the summer of 1993 were used for the
greenhouse and incubator experiments.
Field experiment
An experiment employing a randomized complete block design with five
replications was conducted in 1993 and repeated in 1994 at the Purdue University
Agronomy Research Center in West Lafayette, Indiana. The experiment included one
hundred recombinant inbred lines and five other cultivars as checks. A single row plot, 5
m long and 75 cm between rows, was used. Each plot was drill seeded and hand-thinned
to a spacing of six plants per foot.
The entries were visually scored for seedling vigor on a scale of 1 (most vigorous)
to 9 (least vigorous). The scores were assigned at seedling stage (2-3 leaves), when the

---

differences between entries were observed to be large; this stage corresponded to 29 and
15 days afier planting respectively for the 1993 and 1994 experiments. The visual scoring
system was a relative evaluation based on the range of varïation of seedling size in the
population under study . The estimates visually integrate emergence percentage, plant
height, thickness of leaf canopy and the length and width of individual leaves.
Greenhouse experiment
The experiment was conducted on a sand bench during June-JuPy 1995. The
experiment included ninety-nine recombinant inbred lines and the two parents, with Iine
number two not included because of inadequate amount of seeds. Additionally data for
lines five #and six were not included in the analysis because of mixture. A t-andomized
complete block design with three replications was employed. A single row plot, 30 cm
long and 10 cm between rows was used. Fifty kernels per row were placed at 2 cm depth
and watered daily. One week aRer planting, the number of emerged plants per row was
recorded. The first seedling height was also measured at this time, and each plot was
thinned to 20 plants. Seedling height was again taken at 2 and 3 weeks after planting.
These measurements were taken from the soi1 surface to the tip of the leaves. Ten
seedlings were harvested at 2 and 3 weeks aRer planting, dried at 120” C for 5 days, and
shoot dry weight was recorded.
Three growth rates were calculated from the seedling height and shoot dry weight
measurements. The first (GRl) and second (GR2) growth rates were obtained respectively
between seedling height 1 and seedling height 2 and between seedling height 2 and

---

40
seedling height 3. Similarly growth rate 3 (GR3) were obtained from shoot dry weight 1
and shoot dry weight 2.
Laboratory experiments
Two experiments with two replications each were conducted in an incubator. The
temperature in the incubator was 12’ C for the first experiment and 22” C for the second,
both were at 100 % relative humidity. Fifty seeds of each entry were placed on a filter
paper in a 100 x 15 mm (diameter by height) petri dish, and moistened once every two
days. Eaçh experiment included one hundred recombinant inbred lines and their two
parents. Qne week afier the beginning of each experiment., the seeds were placed at
-70’ C to stop a11 physiological processes. The number ofgerminated seeds was
subsequently recorded as determined by radicle protrusion through the seed coat, in
accordance with the Association of Officia1 Seed Analysts (AOSA, 1970) definition of
germination.
Data analysis
The field visual scores were used to determine inheritance and the number of genes
controlling seedling vigor in sorghum. At the F7 generation of a line derived through the
Single Seed Descent method, nearly a11 loti are expected to be homozygous. We
hypothesized a duplicate or triplicate dominant model, in which the presence of at least
one dominant homozygous locus of either gene resulted in a vigorous phenotype. With a
duglicate model, three genotypic classes with at least one homozygous dominant locus are

---

41
expected, to give a vigorous phenotype and one with both loti homozygous recessive and
non-vigorous plants. It would be expected that the 100 recombinant inbred lines
population be divided into 75 vigorous and 25 non-vigorous lines. A triplicate mode1
would give seven classes with at least one homozygous dominant locus and one with a11
loti homozygous recessive. It would then be expected that, the seven classes would result
in 87.5 vigorous lines, while the remaining class would consist of 12.5 non-vigorous
genotypes. Chi-square analyses were computed to test the goodness-of-fit of the data to
the above hypothesized genetic ratios. Heritability estimates were calculated for the
different variables measured in the field, greenhouse and laboratory experiments. The
mean squares for family [MS(fam)] and for the experimental error [MS(error)] were
obtained fiom an analysis of variante and were used to obtain the genotypic variante. The
additive variante was calculated by adjusting for inbreeding using the formula 02A = 2 -
(U2) t-l where t is the generation of evaluation, SO that at F7, the genotypic variante
contains 3 1/16 of the additive variante in the F2 generation. Broad-sense heritability on a
family mean basis was calculated as follows:
Hf = MS(among lines)- MS(YxL) J MS(among lines) for field traits.
Hr = MS(among lines)- MS(error) / MS(among lines) for greenhouse and incubator
experiments.
The analyses of variante and covariance provided the sum of squares and cross
products SO that the genetic correlations could be obtained, using the formula;
rg = @F(xy)) / @F(xx) . “P(n))1’2.

---

Where: rg = the genetic cor-relation between x and y.
oF(xy) = family covariance component between x and y.
aFtmj = family variante component for variable x
o’F(yy) = family variante component for variable y.
The analysis of covariance between traits measured in the ,field and controlled
environments was calculated assuming that the expected environmental covariance was
zero, SO that the mean cross product for family MCP (fam.) = cr2~cmJ
Standard errors for estimates of genetic cor-relations were calculated according to
the procedure given in Mode and Robinson (1959) using the computer program
SPHENGE (Santini, Nyquist et al., 1992 at Purdue University). Significance at the a =
0.05 and 0.01 levels for genetic correlations was declared if the coefficient exceeded its
standard error by two and three times respectively.

---

43
RESULTS AND DISCUSSION
Highly significant differences for seedling visual scores between genotypes were
obser-ved (Table 1.1). The performance of the sorghum lines for the visual scores was
fairly consistent from year to year. A given line tended to have the same phenotype
(vigorous or non-vigorous ) in both years, as shown by the high cor-relation between 1993
and 1994 scores (R = 0.64, P = 0.001). Visual scoring has been found to be an effective
method for estimating seedling vigor (Maiti et al., 1981). However, they considered the
relative nature of the scores to be an obvious limitation of the method that would not
allow comparisons between experiments or generations of breeding materials. The highly
significant correlation of 1993 and 1994 scores in our study confirm the reliability of the
method for evaluating and comparing experiments and therefore contradicts the latter
statement by Maiti et al. (198 1).
The differences in seedling vigor between the two parents were clearly reflected in
their visual scores. Shanqui Red had a mean score of 1.8 and 1 respectively in 1993 and
1994, while the corresponding scores of SRN39 were of 7.9 and 7 (Table 1.2). Among the
recombinant inbreds, lines with a seedling vigor score between 1 and 4.5 were considered
vigorous, while those with a rating of 4.6 to 9 were regarded as non-vigorous.
Considerable variation in the scores among the different inbred lines were observed.
In
1993, the mean scores (average of five observations) among inbred lines ranged from 1.6
to 7.4 (Fig. 1. l), while in 1994 the mean values were between 1.0 and 9.0 (Fig. 1.2). In
both years the frequency distribution was skewed toward the vigorous parent, Shanqui

---

45
The significant additive genetic variante observed for the visual scores and for the
different measures of seedling vigor in controlled environments, indicate that there is
adequate variation to allow improvement for germination under cold temperature, for
germination and emergence at normal temperatures, and for seedling growth and
development. Cold tolerance is enhanced by rapid germination and high percentage
germination. The parental line Shanqui red originated in a temperate environment in
China, and we hypothesize that the genes controlling seedling vigor may also be valuable
for cold tolerance. The germination percentage of Shanqui red at 12” C was high (87 %)
while that of SRN39 was low (24 %). The significant genetic variante and high
heritability of percentage of germination at low temperature indicate that Shanqui red cari
be effrciently used in breeding programs for improving stand establishment of grain
sorghum in environments where low soi1 temperatures prevail at planting time.
McConnell et al. (1979) improved cold germination at 7.2” C in maize afier four cycles by
about 9 %. A 15 % increase in cold germination was obtained afier four cycles of
phenotypic recurrent selection in sorghum, under early spring planting (Bacon, 1986). The
high heritability estimates for germination at 22” C, as well as for emergence and seedling
height in our study also indicate that attempts to improve these traits would be successfùl.
But progress in breeding for improved seedling dry weight is expected to be slow as
suggested by the low to moderate heritability estimate.
Estimates of genetic correlation among different measures of seedling vigor in the
greenhouse experiment indicated significant positive relationhips between percent
emergence and seedling dry weight at 2 weeks alter planting, and between seedling height

---

46
and seedling dry weight (Table 1.8). The high genetic cor-relation between seedling height
and seedling dry weight indicate that efficient selection for seedling height would improve
seedling dry weight. Genetic correlation coefficients between emergence and seedling
height and weight were relatively small, suggesting that a simultaneous selection scheme
could be used to achieve both high percentage emergence and taller / heavier seedlings.
Germination at 12” C was significantly associated only with shoot dry weight 1 (rg = .49)
and with none of the other estimates of seedling vigor in the greenhouse. Germination at
22”’ C was highly significantly correlated with percent emergence (rg = 1.04) and was also
associated with seedling height 1, II and III and shoot dry weight 1 and II and growth rate
1. No association was observed between germination at 12” C and germination at 22” C
(Table 1.9). Significant associations were revealed between the visual field seedling vigor
scores and germination at 22” C, as well as with emergence, seedling height, shoot dry
weight and growth rate 1 (table 1.10). Germination at 12” C was weakly but significantly
related with the seedling visual scores (fg = -0.12). (The negative sign in the genetic
correlations, results irom the fact that a score of 1 designated the vigorous phenotypes and
and a score 9 represented the non-vigorous lines). There were high genetic cor-relation
between visual scores and percent germination at 22” C and and between emergence,
similarly significant genetic cor-relation was also found between gerfination at 22O C and
emergence. These results suggested that these two traits are controlled by the same
gene(s) in thés population. The non significant genetic correlation between percent
germination at 12” C and germination at 22’ C and emergence suggests however, that
germination under cold temperature is controlled by a different set of genes than those

---

47
controlling germination and emergence at normal temperatures. These results may exnlain
the non significant association between seedling vigor estimated as visual scores in the
field with germination at 12O C. McConnell et al. (1979) made the same observation, and
attributed the lack of cor-relation between laboratory cold emergence and field seedling
vigor results to the difference in conditions between laboratory and the mild spring during
the two years of evaluation. Mock and Eberhart (1972) also found environmental
influences to be large when selecting for cold tolerance. They later reported that
improved seedling vigor accompanied cold tolerance (Mock et al., 1976). Kernel weight
was poorly associated only with percent emergence, and with none of the other estimates
of seedling vigor (table 1.9). This was in agreement with previous observations
(Vanderlip et al., 1973). Others studies have concluded that kernel weight has little effect
on stand establishment (Radford et al., 1990; Mian et al. 1992).
Our results indicate that the visual scoring system used was effective in integrating
germination and emergence at hight temperatures, seedling height, shoot dry weight and
growth rate 1. This indicates that the visual scoring system cari be efficiently used in a
breeding program to improve seedling vigor. There is an apparent limitation to the
improvement of seedling cold tolerance, due to the dependence on direct selection under
the prevailing environmental conditions (McConnell et al., 1979; Mo& et al., 1979). This
cari be overcome with the identification of molecular markers associated with seedling
vigor and their use in breeding programs. Molecular markers, unlike phenotype-based
assays, are not a function of the environment, which cari confound the expression of a
genic trait (Stuber, 1992). Sorghum cultivars with a satisfactory level of cold tolerance

---

could eventually be suitable for early spring planting, when low soi1 temperatures restrict
germination and stand establishment of intolerant genotypes.

---

49
Table 1.1. Expected mean squares and analysis of variante for seedling vigor on
visual scores combined across years.
Source
D F
Expected mean squares
Mean squares
Year (Y) y-l = 1
oc2 + 1ci2, + rGyt + rloy2
726.785294””
Rep. l Y
y (r-l) = 8
9.936274**
Entries (E)
L-l
= 1 0 1
17.480072””
P vs L
1
37.612254””
Between P P-l
= 1
180.00000””
Among L L-l
=
9 9
15.635101""
YxE
(y-l) (L-l) = 101
3.704106""
Y x PVSL
1
4.0237""
YXP
(y-l) (P-l) =l
0.0000"
Y x 1,
L - l =
9 9
3.3783""
Error
y (r-l) (L-l) = 808
0.69

---

Table 1.2. Performance of parental and recombinant inbred lines for seedling
vigor traits.
Parental
Recombinant inbreds
Traits
sRN39
SQR
Mean
Range
mean
mean
93 scores
7.8
1.8
3.87 f 0.07
1.6-8.4
94scores
7
1
2.16 * 0.07
1 - 9
Emerg. (?h)
58
97
69.64 f 1.65
2 9 - 1 0 0
Heightl (cm)
9.3
1 6
13.08 f 0.16
9.3-17
Height2 (cm)
17.3
30.3
27.73 f 0.34
17-34
Height3 (cm)
28.3
4 1 . 6
37.13 f 0.39
26-45.6
Weight 1 (g)
2 4 4 . 6
375
339.27 rt15.28
185-443
Weight2 (g)
6 3 0 . 6
1030.6
857.96 rt153.18
4 9 3 - 1 1 8 2
seed weight (g)
3.475
2 . 6 5 0
2.86 -t 0.04
1.97-4.04
Growth rate 1
0 . 9
1.8
1.9 f 0.04
0.9-2.65
Growth rate II
1.35
1.65
1.21 & 0.03
0.6-2.1
Growth rate III
4 4 . 3 5
71.05
82.44 dz 17.46
23.8-1783
Germ. 12°C (%)
2 4
87
59.66 zt 2.24
9.5-96
Germ22”C (%)
4 3 . 2 5
98
74.63 -t 1.24
43.25-98

---

Table 1.3. Chi-square (x’j analyses of seediing vigor scores for two genetic models.
Expected number of lines
Observed number of lines
Vigorous
Non-vigorous
Genetic mode1
Vigorous Non-vigorous
1 9 9 3 1 9 9 4 mean
1993 1994
mean
x2
P
Duplicate dominant 75
25
7 4
2 6
0 . 0 5 3 o-9-0.7
Triplicate dominant
87.5
12.5
9 1
9
1.12
0.5-0.2
87
1 3
0.023 0.9-0.7

---

52
Table 1.4. Analyses of variante for seedling vigor traits measured in the greenhouse.
-
Emergence
Seedling height 1 Seedling height II
Source
DF MS
M S
M S
Reps (r)
1
0.020202
3.1565
1612.2475
Entries
98
527.610390**
5.9128""
30.8371""
Par. vs Lines
1
227.78726"
0.9597"
77.5595**
Bet. Parents 1
1521.0000””
49.0000**
169.0000""
Among Lines 9 6
520.385739**
5.5156""
28.9113""
Error
98
63.04061
1.5749
8.1352
Seedling. height III
Shoot weight 1
Shoot weight II
Source
DF MS
M S
M S
Reps (r)
1
3033.4595
85147.6566
2602704.0455
Entries
98
37.9591""
7965.9499**
62693.0103**
Par. vs Lines
1
22.2268"
10018.3976"
40271.9870"
Bet. Parents 1
225.0000**
27060.25""
123552.2500**
Among Lines 9 6
36.1747**
7745.6713””
62292.6122””
Error
98
13.1494
4805.7382
33120.9332

---

Table 1.4 cent. Analyses of variante for seedling vigor traits measured in the
greenhouse
Growth rate 1
Growth rate 11
Growth rate III
Source
D F
M S
M S
M S
-
-
Reps 6)
1
3 0 . 0 5 5 6
4 . 5 4 5 4
3 5 6 3 9 . 2 8 2 0
Entries
98
0.26””
0.2491”
782.3607”
Par. vs Lines
1
1.2503””
0.3418”
206.4852”
Bet. Parents 1
0.7347””
0.0816”
713.6645”
Among Lines 9 6
0.2441””
0.2491”
782.3607”
Errer
9 8
0.1133
0.23 1 9
5 3 0 . 3 4 0 7

---

Table 1.5. Analyses of variante for lOO-kernel weight ad percent germination at 12 “C
and 22” C
lOO-kernel weight germination 12” c
germination 22” C
Source
D F M S
M S
M S
.Reps (r)
1
0 . 7 4 6 4
706.4444
18.7929
Entries
98
0.2698””
1066.7518’**
335.1590**
Par. vs Lines
1
0.1519**
51.1716”
46.6761”
Bet. Parents 1
0.6806””
3969.0000**
3080.2500”
Among Lines 9 6
0.2667””
1066.7518**
309.5694””
Error
98
0.2316
353.4240
102.9766

---

55
Table 1.6. Genetic, additive and environmental variante estimates of seedling vigor traits.
Trait
G2G
02*
G2E
--
Visual scores
1.2256
0.6326
0.6947
o/o Germination at 12”~
356.6639
184.0846
353.4240
% Germination at 22”~
103.2964
53.3 127
102.7660
Emergence
228.6725
118.0245
63.0406
Seedling height 1
1.9703
1.01169
1.5749
Seedling height 2
10.3880
5.3616
8.1352
Seedling height 3
11.5126
5.9420
13.1494
Shoot dry weight 1
1469.9665
7 5 8 . 6 9 2 4
4805.7382
Shoot dry weight 2
14585.8395
7 5 2 8 . 1 5 5 2
33 120.9332
1 OO-seed weight
0.1218
0.0628
0.023 12
Growth rate 1
0.0654
0 . 0 3 3 7
0.1133
Growth rate II
0 . 0 0 8 6
0 . 0 0 4 4
0.23 19
Growth rate III
126.0100
6 5 . 0 3 7 4
530.3407

---

56
Table 1.7. Estimates of broad-sense heritability for seedling vigor traits
of F7 recombinant inbred Family means.
Heritability
Environment
Visual scores
0.7839
Field
lOO-seed weight
.9121
Field
% Emergence
0.8774
Greenhouse
Seedling height 1
0.7144
Greenhouse
Seedling height 2
0.6089
Greenhouse
Seedling height 3
0.3670
Greenhouse
Seedling dry weight 1
0 . 3 9 7 9
Greenhouse
Seedling dry weight 2
0.45780
Greenhouse
Growth rate 1
.5274
Greenhouse
Growth rate II
.0796
Greenhouse
Growth rate III
.3219
Greenhouse
% germination at 12” c
0.6565
Incubator
% germination at 22” c
0.6604
Incubator
.c

---

-
57

---

Table 1.9. Genetic correlation coefficients and their standard errors (in parenthesis) of
seedling vigor traits measured in the greenhouse and incubator.
Traits
Germination 12” C
Germination 22” C
lOO-kernel weight
Emergence
.0544 (. 1358)
1.0376** (.0696)
-.2407* (. 1084;)
Seedling height 1
.0939 (. 1558)
.3689** (. 1481)
.2019 (. 1195)
Seedling height II
.0452 (.1543)
.3067+ (. 1465)
1904 (. 1196)
Seedling height III
.1532 (1683)
.2289* (. 1636)
.0722 (. 1375)
Shoot dry weight 1
.4992* (.2411)
.8210** (2193)
.3525 (.1782)
Shoot dry weight II
.2286 (.2008)
.3872*(.1890)
.2526 (. 1575)
Growth rate 1
.0079 (.1856)
.2738* (. 1774)
-.9150 (1.0727)
Growth rate II
.5345(.7878)
-.33 14(.6556)
-.9150(1.0727)
Growth rate III
.1080 (.2442)
-. 1939 (.2403)
.2144 (. 1937)
Germination 12” c
.2641(. 1591)
-.1620 (l303)
Germination 22” c
-. 1166 (-1302j

---

59
Table 1.10. Genetic correlation coefficients and standard errors (in parenthesis) of
seedling visual scores, and traits measured in greenhouse and incubator.
--
Traits
Scores
1 OO-seed weight
.1078 (x222)
Emergence
-.5595** (.0956)
Seedling height 1
-.3530** (1330)
Seedling height II
-.4398** (.1294)
Seedling height III
-.5380** (.1424)
Shoot dry weight 1
-.6717** (.2047)
Shoot dry weight II
-.4082* (. 1826)
Growth rate 1
-.5220** (. 1588)
Growth rate II
-.5328 (.7752)
Growth rate III
.2676 (.23 19)
Germination at 12” c
-1224 (. 1490)
Germination at 22” c
-.6003** (. 1140)

---

PERCENTAGE OF INBRED LINES
3

---

50
45
g 40
5
w 35
m 30
Z
8 25
z
2 20
g 15
5
[1 10
5
0
1
2
3
4
5
6
7
8
9
VISUAL SCORES
Figure 1.2. 1994 inbred line distribution for seedling vigor scores.

---

62
CHAPTER T W O
RELATIONSHIP OF SEEDLING VIGOR
ESTIMATES WITH AGRONOMIC
TRAITS IN SORGHUM

---

63
ABSTRACT
Crop yield is a mnction of plant density, the efficiency of dry matter production,
and the proportion of the biomass that is economic yield. Germination and emergence cari
influence plant population density. However the direct eEects of seedling vigor traits on
su.bsequent plant performance are more difficult to discern. The objective of this study
was to evaluate the genetic relationships of seedling vigor traits with traits in later stages
of plant development in sorghum (Sorghum bicolor L. Moench). Field and controlled-
environment experiments were conducted to obtain estimates of seedling vigor
parameters, grain yield, plant height, and maturity in a recombinant inbred population.
Germination at 22” C, emergence, seedling vigor scores and the rate of dry matter
accumulation at seedling stage were significantly correlated with grain yield. However no
association was found between kernel weight, shoot height and grain yield. The
association of early maturity with high percent germination and emergence, shoot growth
(height and weight) and the rate of dry matter accumulation was believed to result from
tight linkage in the parental line ‘Shanqui red’. It was concluded that this cultivar could be
a valuable parent for the development of early maturing varieties, with improved stand
establishment, and adapted to environments where low ternperatures prevail at planting in
spring and early frost in the fall.

---

64
INTRODUCTION
Seedling vigor cari be measured by many variables including kernel weight,
germination, emergence and seedling growth (shoot height, dry weight and growth rate).
Many studies emphasized the relationship of laboratory germination and vigor to tield
emergence. There is less published information relating seedling vigor to other aspects of
trop performance (Tekrony et al., 1991). However there are possibilities that seedling
vigor cari influence trop yield (Ellis, 1992). Differences in germination and emergence cari
influence plant population density. Since trop yield is a function of density (Willey et al.,
1969) it follows that seedling vigor cari influence trop yield through germination and
emergence. If seed quality (size, percent germination) only affected percent emergence,
then growers could, theoretically, overcome such effects by adjusting seed sowing rate. It
ha.s been shown that the effect of differences in laboratory germination on field emergence
in different seedbeds cari be quantified (Khah et al., 1986; Wheeler et al., 1992).
However, in practice adjustments in seed sowing rates are hampered by difficulties in
forecasting the particular seedbed environment (Ellis, 1992).
Seedling vigor also affects trop performance through effects on the plant growth
processes involved in the production of yield. Yield of any trop is determined by the
insolation intercepted by the plant community, the efficiency with which intercepted
insolation is converted to dry matter, and the proportion of the biomass that is economic
yield (Charles-Edwards, 1982). Most of the plant tissues involved in the production of
dry matter and yield are formed afier seedling emergence, and it seems unlikely that seed

---

65
quality (weight, percent germination) would influence their ability to carry out
physiological processes and accumulate dry matter. Seed quality did not affect the relative
growth rate of soybean seedlings, provided they were free of physiological injury or
necrotic lesions (Egli et al., 1990). Priming or natural variation in seed quality bave been
reported ‘to have no effect on the relative growth rate of onions (Allium cepa L.; Ellis,
1989). However genetic aberrations which may occur in long term storage could impair
physiological function in later formed plant tissue (Roberts, 1972; Harrison, 1966).
Seedling vigor may have a direct effect on the ability of the plant to accumulate
dry matter. The direct effects of seedling vigor on yield greatly depend upon the trop
species. Crops harvested during vegetative growth are fi-equently planted at low
population densities and harvested on an individual plant or area basis as vegetative mass
of aboveground [Lettuce (Lactuca sativa L.), cabbage (Brassica oleracea var capitata)]
or underground [sugarbeet (Beta vdgaris L.), radish (Rajbms sativus L., carrot (Daucus
carota L.)] structures. The effects of seedling vigor cari be specially critical in these
crops, where delayed emergence or missing plants may reduce yield and uniformity at
harvest (Tekrony et al., 1991).
Crops harvested at an early stage of reproductive development usually are planted
at higher population densities than crops harvested during vegetative growth. The
quantity (fresh weight), uniformity, and quality of reproductive structures determine the
yield. Perry (1969) evaluated high and low-vigor seed lots of two pea cultivars and
reported significantly lower yield for low-vigor seed lots thinned to similar plant
population as high-vigor lots. Abdalla and Roberts (1969) stored pea seeds in various

---

66
artificial environments, and reported that seed lots with lower viability had lower early
plant growth rates (O-35 days alter planting) and reduced plant height and leaf number at
45 and 59 days after planting. At later stages of growth, no significant differences in
relative growth rate were recorded, and reductions in dry seed yield occurred only for
those seedlots where viability had declined below 50%. LJnfortunately, measurements of
fresh seed and fruit weight at an early stage of reproductive development were not taken.
A non significant relationship between germination and fresh pod weight was reported in
lima beans (Phaseolus Zinensis L.) (Bennett and Waters, 1984).
Grain crops are harvested alter they have completed their life cycle (full
reproductive maturity) and only the seeds are harvested for yield. Work with soybean
(Tekrony et al., 1987), bean (Spilde, 1987), and corn (Abegbuyi et al., 1989) suggests that
there is no relationship between germination, seedling dry weight and seedling growth rate
with yield. However significant increases in yield with seedling growth rate were shown
for corn (Burris, 1975). Sign&ant increases in yield with high percent germination of
seed were also shown for corn when grown at low population densities compared with
seed that had been stored for 5 to 7 years (Funk et al., 1962). It was also demonstrated
that seedling growth rate of spring barley showed an advantage only at lower plant
density, while no association existed at normal population density (Perry, 1980). It was
found that low percent-germination seeds of spring wheat produced lower yields only in
lower than normal populations or later than normal plantings. Yield of soybean was
reported to be related to accelerated aging, cold test, seedling and seedling growth rate in
hi11 plots but not in row plots planted at normal populations. In sorghum a significant

---

67
relationship between grain yield and speed of germination, root and shoot growth was
reported (Camargo et al., 1973); however the plant density (66, 000 plants / ha) was
lower than normally recommended (Vanderlip, 1972). No significant association was
found bet.ween kernel weight and days to 50 % bloom, plant height and grain yield in
sorghum (Suh et al., 1974). It was also reported that kernel weight had no influence on
grain yield of sorghum (Maranville et al., 1977).
From these studies, it was concluded that a yield response to seedling vigor traits
occurs only when plant densities are lower than the density required to maximize yield.
However no report is available in sorghum on the genetic correlation of seedling vigor
variables with the performance of the trop. The objective of this study was to evaluate
the genetic relationships of seedling vigor traits with yield, plant height and maturity at
high plant density recommended for commercial production.

---

68
MATERIALS AND METHODS
Recombinant inbred (RI) lines of a cross between ‘SRN39’ and ‘Shanqui red’
(SQR) were obtained through the Single Seed Descent method of plant breeding. These
two parents differ in seedling vigor with Shanqui red being the more vigorous. An
experiment employing a randomized complete block design with five replications was
conducted in 1993 and repeated in 1994 at the Purdue University Agronomy Research
Center in West Lafayette, Indiana. The experimental materials and designs are the same as
described in chapter one. The entries were visually scored for seedling vigor on a scale of
1 (most vigorous) to 9 (least vigorous). The number of days from planting until anthers
had extruded half way down the panicles of at least 50 % of plants in a row was used as an
estimate of maturity. Plant height was measured centimeters from the soi1 surface to the
top of‘the panicles alter maturity. Grain yield (kg / ha) was estimated from the tenter 3 m
of a single row plot in 1993, and in 1.994 from the entire 5 m of a single row plot. Grain
samples taken from replications 1 and III at harvest in 1994 were used to estimate lOO-
seed weight. Because of the high heritability of grain size (0.91) this seed-weight value
was also used as an estimate of kernel size at planting.
The laboratory experiments conducted in an incubator during February-March and
the greenhouse experiment conducted on a sand bench during June-July 1995 had the
same entries, design and procedures for data collections as described in chapter 1.
Heritability estimates were calculated for the visual scores, grain yield, plant height
and maturity. These estimates were also calculated for the different variables measured in

---

69
the greenhouse and laboratory experiments. The mean squares for family [MS(fam)] and
for the experimental error [MS(error)] were obtained from an analysis of variante and
were used to obtain the genotypic variante. The additive variante was calculated by
adjusting for inbreeding using the formula cr2* = 2 - (1/2) t-1 where t is the generation of
evaluation, SO that at F7, the genotypic variante contains 3 1/16 of the additive variante in
the F2 generation. Broad-sense heritability on a family mean basis was calculated as
follows:
Ht = MS(among lines)- 1/2 MS(YxL) / MS(among lines) for traits measured in the field.
Hf = MS(among lines)- MS(error) / MS(among lines) for greenhouse and incubator
experiments.
The analyses of variante and covariance provided the sum of squares and cross
products SO that the genetic cor-relations could be obtained, using the formula;
rg = @F(xy)) / @F(@ . ~2F(n))1’2.
Where: rg = the genetic correlation between x and y.
(TF&) = family covariance component between x and y.
o’r(=) = family variante component for variable x.
02F(yy) = family variante component for variable y.
The analysis of covariance between traits measured in the field and controlled
environments was calculated assuming that the expected environmental covariance was
zero, SO that the mean cross product for family MCP (fam..) = 02F@)
Standard errors for estimates of genetic correlations were calculated according to
the procedure given in Mode and Robinson (1959) using the computer program
. _ -

.-
..“I-.m”-..“.,-
--

---

70
SPHENCJE (Santini, Nyquist et al., 1992 at Purdue University). Significance at the a =:
0.05 and 0.01 levels for genetic correlations was declared if the coefficient exceeded its
standard error by two and three times respectively.

---

71
RESULTS AND DISCUSSION
Significant differences were found with a11 the traits (visual scores, 100~seed
weight, plant height, maturity and grain yield) measured in the field (Table 2.1). The
parental line Shanqui red had higher yield, and was earlier and taller than SRN39 (Table
2.2). The recombinant inbred population tended to have intermediary values for these
traits except for 1994 yield when it was lower.
Seedling vigor scores were highly correlated with grain yield and plant height, but
the association with days to maturity was not significant (Table 2.3). Acevedo et al.
(199 1) also found a highly significant cor-relation coefficient between seedling vigor scores
and grain yield in barley. It was also reported in maize that improved seedling vigor at
low temperatures was accompanied by lower plant height (Mock and Bakri, 1976).
Kernel weight was not associated with yield or maturity. This is in agreement with
observations made by Suh et al., (1974); Vanderlip et al. (1973); Maranville et al. (1977).
The significant association of kernel size with plant height is also in agreement with a
previous finding (Ibrahim et al., 1985) but is contrary to another report in sorghum (Suh
et al., 1974).
Germination at 22” C and emergence were significantly associated with yield
(Table 2.4). Such associations cari arise because yield is a function of plant density (Ellis,
1992); and because percent germination and emergence c,an influence plant population.
This result is contrary to earlier reports that percent germination and emergence have no
effect on yield at high density in sorghum (Vanderlip, 1972) and other crops (Tekrony et
.”
.^^““.._

-,s-ir--s.*ml*LIIUII

---

72
al., 199 1). Percent germination at 22” C and emergence were also significantly associated
with plant height, and negatively correlated with maturity. These genetic correlations may
result fi-om the associations of these traits in the parental line, ‘Shanqui red’, which has
higher percent germination and emergence, and is taller and earlier than SRN39. The
single cross between the parental lines, followed by selfing, may not have allowed
sufficient breakage of linkage blocks (Hanson, 1959).
No genetic association was revealed between germination at 12” C with yield,
plant height and maturity. It was also reported that separate genetic systems control cold
tolerance and maturity in corn, allowing selection within a.dapted material without mating
“C
to unadapted and early types (Mock and Eberhart, 1972). However following selection
for germination at 14’ C in sorghum, an increase in grain yield of 198 kg / ha was
observed, plant height showed an irregular response, and days to half bloom decreased by
0.55 days per cycle of phenotypic recurrent selection (Bac:on et al., 1986). Selection in
cor-n for phenotypes which germinated and grew at temperatures below 10” C, gave lines
which flowered in 60 days (Brown, ‘1968). Mock and Bakri (1976) found that improved
seedling vigor , lower plant height, lower harvest moisture and early flowering
accompanied increased cold tolerance.
Shoot growth (height and dry weight) was not associated with yield and plant
height, in agreement with observations made in corn (Adegbuyi et al., 1989) and would
support the statement that in situations where vegetatitve growth is adequate to maximize
yield, it is unlikely that the effect of seedling vigor on vegetative growth Will carry over to
affect yield (Tekrony et al., 1991). Camargo et al., (1973) obtained different results and

---

73
reported a significant relationship between sorghum yield and shoot grooth, but in lower
than normally recommended plant density. In small grain cereals, differences in mass
established early were maintained until physiological maturity and contributed to the better
performance of barley compared to oats and triticale (Lopez-Castaneda ‘et al., 1994b).
Shoot growth (height and dry weight) was also not associated with plant height, but a
negative genetic cor-relation was observed with maturity (Table 2.4).
Growth rates 1 and II (rate of height increase) were not associated with yield, plant
height and maturity. The rate of shoot dry matter accumulation (growth rate III) was
significantly correlated with yield and maturity. Conflicting results have been reported
with different crops. Burris (1975) reported a positive association between seedling
growth rate and grain yield in corn, while no correlation was found between these two
traits in another study in corn (Tekrony et al., 1989) and in bean (Spidle, 1987). The
result of our study suggest, then, that the rate of shoot dry matter accumulation is related
to grain yield at high plant density in sorghum. Growth rate III was also negatively
associated with maturity, as were shoot growth (height and weight), germination and
emergence. Findings by Allard (1988) state that natural selection favors the development
of multilocus clusters conferring adaptation. Linkage between seedling vigor traits and
early maturity in Shanqui red would favor adaptation in its original environment (North
China) w:here low temperatures at planting and early frost are likely.
The significant genotypic and additive variante of germination, emergence, shoot
growth and maturity (Table 2.5) indicate that Shanqui red could be a valuable source
material for the development of early maturing varieties with improved stand

---

establishrnent, and adapted to environments where low temperature prevails at planting in
spring and early Frost in the fall. The high heritability estimates (Table 2.6) of maturity
and seedling vigor traits suggest that such a breeding objective is feasible. However,
separate genetic systems control maturity and grain yield, as indicated by a nonsignificant
genetic correlation (Table 2.3). This makes it a unique population, because yield and
maturity are correlated in most others (Ibrahim et al., 1985; Bacon et al., 1986)

---

75
Table 2.1. Expected mean squares and analysis of variante for seedling vigor visual
scores combined across years.
Source
D F
Expected mean squares
Mean squares
Year (Y)
y-l
=
1
os2 + 10~~ + rq12 + rloy
726.785294**
Rep. / Y
y (r-l) = 8
Oc2 + 1or
9.936274””
Entries (E)
L-l
= 1 0 1
17.480072””
Par. vs L
1
37.612254””
Bet. Par.
P-l
=l
180.00000””
Among L
L-l
=
99
15.635101””
YxE
(y-l) (L-l) = 101
3.704106””
Y x PVSL
1
4.0237””
YXP
(y-l) (P-l) =l
G2 + @YP
~0.0000”
Y x I.,
L - l =
9 9
(JE 2 + ra,12
3.3783””
Error
y (r-l) (L-l) = 808
0 . 6 9

---

76
Table 2.1 Cont. Expected mean squares and analysis of variante combined across years
for grain yield, plant height, maturity and 1 OO-kernel weight
Source
DF
M S
M S
M S
M S
Grain yield
Plant height
Maturity
1 OO-kernel w
Year (Y) 1
169167170.26
301.9853
33185.4157
Rep. /Y
8
3907293.53
793.6544
42.0603
0 . 7 5 6 4
L
Entries (E)
1 0 1 7 4 6 9 5 3 1 . 4 8 0 ” ” 9958.7045**
123.2014** 0.2698**
PVSL
1
237.466321”
224.803””
1.8722””
0.8326””
Between P 1
24557712.20”
5746.0500”” 884.4500** 0.0000”
AmongL 9 9
7132508.44””
10099.578”” 116.7376”” 0.2667
YxE
101 3753431.980”” 279.3912**
12.7642
Y XPVSL 1
3 4 4 6 8 7 6 . 2
46.2607””
0.4833”
YXP
1
10512500.00”” 42.0500””
6.0500””
Y x L
9 9
3688255.090**
284.1435**
12.9561**
Error
808 1897631.80
102.25
4.1663
0.0234

---

77
Table 2.2. Field performance of parental and recombinant inbred lines.
Parental
RI lines
Traits
sRN39
SQR
mean
Range
mean
mean
93 visual scores
7.8
1.8
3.87k 0.07
1.6 - 8.4
94 visual scores
7
1
2.16 f 0.07
l-9
93 yield (kg/ha)
3002
6668
4170 f 107.:34
1825 -6668
94 yield (kg/ha)
6088
6854
4975h98.45
2619-7323
9 3 maturity (d)
93
78
85.32 f 0.40
78-94
94 maturity (d)
80
68
73.83 f 0.33
67-84
9 3 height (cm)
160
196
182.63 f 3.08
107-246
9 4 height (cm)
165
196
183.86% 3.37
107-254

---

78
Table 2.3. Genetic correlation coefficients and their standard errors (in parenthesis)
Among field performance traits.
Traits
Yield
Maturity
1 OO-kernel wght
Vigor
-0.8165** (0.173) -0.2268** (0.113) 0.0322 (0.124) 0.1078 (0.122)
Yield
0.5461** (0.141) 0.2325 (0.161) -0.1735 (0.165)
Height
-0.0954 (0.107)
0,3539** (095)
Maturity
0.1577 (0.109)

---

79
Table 2.4. Genetic correlation coefficients and their standard errors (in parenthesis) of
seedling vigor traits and field performance.
Trait
Yield
Plant height
Maturity
--
Vigor scores
-0.8165** (0.173 1) -0.2268** (0.1130) 0.0322 (0.1248)
1 OO-seed weight
-0.1735 (0.1658)
0.3539** (0.0952) 0.1577 (0.1097)
Emergence
0.5126** (0.1595)
0.1279 (0.1081)
-0.2822* (0.1075)
Shoot height 1
0.0523 (0.1953)
0.1154 (0.1214)
(-0.4392)** (0.1134)
Shoot height II
0.0023 (0.1948)
0.0948 (0.1212)
-0.3741**(0.1158)
Shoot height III
0.0448 (0.2166)
0.1106 (0.1317)
-0.4541** (0.1248)
Shoot. dry weight 1
0.1010 (0.2905)
0.2437 (0.1707)
-0.6131**(0.1755)
Shoot dry weight II
0.1947 (0.2569)
0.1507 (0.1524)
-0.5716** (0.1491)
Growth rate 1
-0.0561 (0.2215)
-0.0641 (0.6079)
0.2476 (0.3059)
Growth rate II
0.0985 (0.1357)
0.0972 (0.3598)
0.1242 (0.1799)
Growth rate III
-0.3594** (. 1330)
-0.4526 (0.5759)
-0.591 o* (0.1994)
Germination 12” C
-0.1565 (0.2098)
0.0047 (0.1272)
-0.1780 (0.1334)
Germination 22” C
0.6718** (0.1990)
0.2194*(0.1217)
-0.2728** (0.1255)
_ __ _ _-
-.--
W.

---

X0
Table 2.5. Genotypicc, additive and environmental variante estimates of seedling vigor
and other field performance traits.
Trait
02G
02*
02E
-Grain yield
344425.335
177767.9148
189763 1.8
Plant height
981.5434
506.60307
102.25
Maturit,y
10.3781
5.3564
4.16
Visual scores
1.2256
0 . 6 3 2 6
.6947
Emergence
228.6725
118.0245
63.0406
Seedling height 1
E .9703
1.0169
1.5749
Seedling height 2
10.3880
5.3616
8.1352
Seedling height 3
11.5126
5.9420
13.1454
Shoot dry weight 1
1465.9665
7 5 8 . 6 9 2 4
4805.7382
Shoot dry weight 2
14585.8395
7 5 2 8 . 1 5 5 2
33 120.9332
1 OO-kernel weight
.1218
.0628
.0216
Growth rate 1
.0654
.0337
.1133
Growth rate II
.0086
.0044
2319
Growth rate III
126.01
65.0374
530.3407
% Germination at 12°C
356.6639
184.0846
353.4240
% Germination at 22°C
103.2964
53.3 127
102.766

---

81
Table 2.6. Broad-sense heritability estimates of seedling vigor and
field performance traits.
Traits
Heritability
Test environment
Visual scores
0 . 7 8 3 9
Field
1 OO-kernel weight
.9121
Field
% Emergence
0.8774
Greenhouse
Seedling height 1
0.7086
Greenhouse
Seedling height 2
0.7144
Greenhouse
Seedling height 3
0.6089
Greenhouse
Seedling dry weight 1
0.3670
Greenhouse
Seedling dry weight 2
0.4578
Greenhouse
Growth rate 1
.5274
Greenhouse
Growth rate II
.0796
Greenhouse
Growth rate III
.3219
Greenhouse
% germination at 12” C
0.6565
Incubator
% germination at 22’ C
0.6673
Incubator
Height
0.9754
Field
Maturity
0 . 8 8 9 0
Field
Yield
0 . 4 8 2 9
Field

---

x2
CHAPTER THREE
EFFECT OF PHENOLIC CONCENTRATIONS
IN SOR.GHUM KERNELS ON SEEDLING VIGOR AND PLANT PRODUCTIVITY

---

83
ABSTFUCT
Sorghum (Sorghum bicolor L. Moench.) plants produce large amounts of phenolic
compounds which constitue a defense mechanism against fùngi , insects, and birds, but
also have antinutritional effects.The role of phenols on seedling vigor and stand
establishment are not well understood. One hundred recombinant inbred (RI) lines
derived fiom two parental lines that differ for an array of phenolic compounds were
evaluated. Highly significant genotypic and additive components of variante, and high
broad-sense heritability estimates were found for pigments, flavan-4-ols, tannin, and total
phenols content of the grain. High phenolic compound contents wet-e associated with
vigorous seedlings, high percent germination at 22” C, emergence, and taller seedlings.
Kernel weight was negatively associated with concentrations of these compounds. Only
total phenols were signifïcantly associated with grain yield. Lines with a red coleoptile,
wbich had signifkantly higher pigment and total phenol contents, tended to be more
vigorous at seedling stages than lines with a green coleoptile. No association of coleoptile
color with plant height and maturity were observed. However, lines with a red coleoptile
tended to be more productive. Because tannin content did not signifkantly contribute to
total phenols (non significant genetic correlation), it was suggested that there are
possibilities to maintain the positive association of total phenols with seedling vigor, and
reduce at the same time reduce the antinutritional effects of tannin in grain sorghum.

---

83
INTRODUCTION
Sorghum plants produce large amounts and a great diversity of phenolic
compounds (Butler, 1989). Many of these phenols determine plant color, appearance,
nutritional quality, and host defenses. Polyphenols are secondary metabolites, their
amount and nature vary greatly with genotypes and environmental conditions under which
plants are grown. Sorghum phenolic compounds cari be divided into five basic groups;
phenolic acids, lignins, quinones, flavonoids and tannin (Butler, 1989). Phenolic acids,
flavonoids and tannins are the most common groups in sorghum. Chemically, phenolic
acids are ,the simplest polyphenols of sorghum. Eight different phenolic acids were
identified in sorghum grains (Hahn et al., 1983). Ferulic acid was the most abundant.
Anthocyanidins are the major pigments in most plants. In sorghum the dominant
pigments are the 3-deoxyanthocyanidins. The color of sorghum grains is influenced by
pericarp color, mesocarp thickness, presence of testa, and by endosperm texture and color
(Hahn and Rooney, 1986). The pericarp color is determined by two genes (Kambal et al.,
1976) and cari be white, lemon-yellow or red. Kambal et al. (1976) did not find any
visible pigments in white-seeded grain but considerable amounts ofp-coumaric, caffeic
and ferulic acids were detected. The pigment in the yellow grain was identified as
eriodictyol chalcone, a deep yellow pigment. The red seeds contained the 3-
deoxyanthocyanidins, luteolinidin and apigeninidin. Doggett (1988) classified sorghum
seedlings in two groups, red and green. Red coleoptile color is controlled by a single
dominant gene over green.

---

85
Flavan-4-01s also called leucoanthocyanidins since they are converted to
anthocyanidins when heated in acid with the loss of a water molecule (Watterson and
Butler, 1983), include monomers of flavanols such as flavan-3,4 diols and flavan-4-01s.
The concentration of flavan-4-01s in sorghum seeds is highly dependent on seed maturity
(Jambunathan et al., 1990). Grain at early stages of maturity (10 and 14 days after
flowering) contained the highest flavan-4-01 concentrations, followed by a drastic decrease
with increased maturity. It was suggested that flavan-4-01s could be degraded, converted
or incorporated into other molecules such as 3-deoxyanthocyanidins or tannins.
Tannins are a group of phenolic compounds found in plants, which convert animal
hides to leather during the tanning process (Butler, 1989). There are two classes of
tannins: hydrolysable and condensed tannins. Only condensed tannins, which are
oligomers of flavan-3-01s have been found in sorghum. These oligomers are now referred
as procyanidins, because the red anthocyanidin pigment cyanidin is released when the
tannin is treated with minera1 acids. Tannins are the most abundant phenolic compounds
extractable from the seed of brown, bird-resistant sorghum (Hahn et al., 1984). Tannins
bind to and precipitate proteins, reducing the nutritional value of the grain. High tannin
sorghums have different kernel structures from other sorgbums (Hahn et al., 1984). High
tarInin sorghums have a prominent pigmented testa located beneath the pericarp. The
pigmented testa is purple or reddish-brown and varies in thickness. The presence of a
pigmented testa is controlled by the complementary Bl and B2 genes. The S gene
controls the presence of pigments and tannins in the epicarp. When S is dominant, more
phenols and tannins are present in the pericarp.
‘-’ - - - - 1 3 1 1 1 )
*I-m---

---

86
It was reported (Chavan et al., 1981) that the percent germination in sorghum
cultivars with high (3.4 %) and low (0.5 Oh) tannin content was the same. However both
root and shoot growth were markedly suppressed in high tannin as compared to low
tannin seedlings. The rates of germination were also the same, but the subsequent rates of
root and and shoot growth were much lower in high tannin seeds. The assayable tannin
content decreased markedly during germination. Tannins are located in the seed coat of
the sorghum grain (Jumbunathan et al., 1973). The loss of tannin was attributed to
leaching in growth medium and penetration into the endosperm with imbibed water during
germinatilon. Starch content decreased, and the rate of formation and total accumulation
ofreducing sugars and free amino ac,ids was lower in high tannin seeds. The interpretation
was that starch and protein degradation were inhibited in high tannin seeds during
germination, leading to suppressed seedling growth. This inhibition would result fi-om the
portion that enters the endosperm. Such tannins are likely to form complexes with seed
protein reserves and enzymes, and inactivate them (Chavan et al., 198 1).
During germination, the reserves of nutrients like starch and proteins are degraded
to soluble sugars and amino acids, respectively, to meet the seedling growth requirements
(Dalvi, 1974). Any depression of starch and protein degradations would indicate
interference with the metabolic systems operating on reserve starch and protein, mainly
enzymes like amylases and proteases. Tannins are reported to form complexes with
hydolytic enzymes and inactivate them (Tamir et al., 1969; Milic et al., 1972). A marked
suppression of seedling root growth was also observed with a low tannin (0.1 Oh) sorghum
cultivar, germinated in the presence 1 %, 2 % and 3 % tannic acid concentrations. The

---

87
inhibit,ion increased with concentration and time (Chukwura et al., 1982). A decrease in
starch content in the control sample (distilled water) and 1 % tannic acid solution, but not
at higher concentrations was also noted. A concommitant increase of soluble
carbohydrate content at low concentration of tannic acid, and in distilled water , and a
decrease at higher concentrations were also observed. The fact that with high
concentrations of tannic acid the soluble carbohydrate falls after germination below its
original level was viewed as an indication that these substances were being utilized, and
that tannic acids directly inhibited their production. Alpha and beta-amylase activity were
also observed to be inhibited by an increase in the concentration of tannic acid. It was
concluded that tannins present in sorghum seeds retard seedling growth due to inhibition
of starch degradation by inactivating hydrolytic enzymes during germination.
In this study we evaluated a recombinant inbred population of sorghum that is
segregating for an array of biochemical and agronomie traits to investigate possible
associations. The specific objectives of this study were; ï) to evaluate the effects of
diEerent phenolic compounds including tannins on seedling vigor and on field
performance; ii) to assess the relationship between the different phenolic compounds and
to estimate their heritability.

---

88
MATERIALS AND METHODS
Recombinant inbred (RI) lines of a cross between ‘SRN39’ and ‘ Shanqui red’
(SQR) were obtained through the Single Seed Descent method of plant breeding. These
two parents differ in seedling vigor with Shanqui red being the most vigorous. An
ex:periment employing a randomized complete block design with five replications was
conducted in 1993 and repeated in 1994 at the Purdue University Agronomy Research
Center in West Lafayette, Indiana. The experimental materials and designs are the same as
described in chapter one. The entries were visually scored for seedling vigor on a scale of
1 (most vigorous) to 9 (least vigorous). The number of days from planting until when
anthers had extruded half way down the heads of at least SO % of plants in a row was used
as an estimate of maturity. Plant height was measured as centimeters from the soi1 surface
to the top of the panicle after maturity. Grain yield (kg / ha) was estimated fi-om tenter 3
m of a plot in 1993, and in 1994 from entire 5 m plot. Grain samples taken from
replications 1 and III at harvest in 1994 were used to estimate lOO-seed weight. Because
of the high heritability of grain size (0.91) this lOO-kernel weight value was also used as
an estimate of kernel size at planting.
Tbe concentrations of the different phenolic compounds were determined from the
grain samples used to estimate lOO-seed weight. The procedures used were as follow:
T&e seeds were ground in a Cyclotec 1093 sample mil1 (Hoganas, Sweden). The
assays were carried out within 2-3 days afier grinding. A sample size of 250 mg was used
for a11 analyses. Methanol containing 0.5 % HCL (15 ml) was utilized as extractant for 20

---

min on a Labquake mixer (LabIndustries, Berkeley, Ca.). Alter centrifigution, the
.-
supernatant was kept on ice until used for analyses. For pigment content, the absorbante
of the 0.5 % HCL / methanol extract was determined at 485 mn. For flavan-4-01s
determination a 0.5 ml aliquot of the 0.5 % HCL in methanol extraction was added to 7 ml
of 30 % HCL in buthanol, incubated at room temperature for 2 hours, and the absorbante
was read at 550 mn. Alter the flavan-4-01s reading, the mixture was heated in a boiling
water bath for two hours. The solution was then cooled for 5 min, and the absorbante
was read at 550 nm to determine the proanthocyanidin concentrations. The blank for
flavan-4-ols and proanthocyanidins was prepared by mixing 0.5ml of 0.5 % HCL in
methanol extract with 7 ml of a mixture of methanol, 0.1 N acetic acid, and
n-buthanol (v/v/v/, 15; 15; 70). Alter two hours, the absorbante was recorded at 550 nm.
Flavan-4-01 and tannin concentrations were determined after correcting for the blanks. Al1
the above absorbantes were read with a spectronic 20 d spectrophotometer (Milton Roy,
Rochester, NY). For total phenols determination a 0.2 ml aliquot of the acidic methanol
extract was diluted with 50 ml of distilled water and 3 ml of 0.05 M FeCL3 in 0.1 N HCL
was added afier 3 min, 3 ml of 0.008 M K3Fe(CN)6 was added and the solution was
incubated again for 19 min. The absorbante of the solution was read at 720 nm with a
Klett Summerson photoelectric colorimeter (Klett Mfg CO. Inc. NY). Total phenols are
reported as Klett units; the other components are reported as absorbante units, a11 at the
wavelengths noted above.
The laboratory experiments conducted in an incubator during February-
March and the greenhouse experiment conducted on a sand bench during June-July 1995

---

had both the same entries, design and procedures for data collections as described in
chapter 1.
The statistical procedures used to estimate varianc-e components, heritability, ‘and
genetic correlations between phenolic compound concentrations with seedling vigor traits
and field performance, were the same as described in chapters 1 and II.

---

91
RESULTS AND DISCUSSION
Highly significant differences were observed for pigments, flavan-4-01s tannin and
total phenol concentrations (Table 3.1). Shanqui red had much higher concentrations of
these compounds than SRN39. Few transgressive segregants with concentrations higher
than those of Shanqui red were found, but no inbred had a level lower than those of
SRN39 (Table 3.2).
The genotypic and additive components of variante for the phenolic compounds
were significant, and their broad-sense heritabilities were high (Table 3.3) and in
agreement with findings by Weerasuriya (1995). Woodruff et al., (1982) reported a
moderate to high broad-sense heritability for tannin quantïty in sorghum, Ma et al., (1978)
also found a high broad-sense heritability for tannin content in common bean.
Significant and positive genetic correlations were found between pigments, flavan-
4-01s tannin and total phenols (Table 3.4). McMillan et., (1972) also found a highly
significant and positive correlation between seed color and tannin content. It was reported
that flavan-401s are converted to anthocyanidins when heated in acid with the loss of ‘a
water molecule (Watterson et al., 1983). It was also suggested that flavan-4-01s could be
degraded, converted or incorporated into other molecules such as tannins (Jambunathan et
al., 1990). These facts may explain the above positive genetic correlations. It was noted
that dark red or brown colored pericarps with pigmented testa were generally higher in
phenolic compounds content than lighter colored varieties of sorghum (Wall et al., 1970).

---

92
Seedling vigor was significantly associated with high tannin concentration. Percent
germination at 22” C and germination, seedling height and its rate of increase were also
highly correlated with tannin concentration (Table 3.5). These results are in disagreement
with tindings by Chavan et al., (198 1); and Chukwura et al., (1982) who observed
Sign&ant reduction in root and shoot growth with increased level of tannin. It was
suggested that tannins present in sorghum seeds retard seedling growth due to inhibition
of starch degradation, by inactivating the hydrolitic enzymes during germination. It was
also reported that extracts from high tannin sericea lespedeza [Lespedeza cuneata (Dum.
cours) G. Don.] residues decreased rye (Secale cereale L.) seed germination whereas
extracts from low-tannin had no effect. Low-tannin residue extract reduced rye coleoptile
length, and a more dramatic reduction was observed with high tannin extract. These
results were explained by the reduction of ce11 elongation or division due to the effect of
tannin (Kalburtji et al., 1993). With two Vicia faba pairs of near-isogenic lines, it was
observed that high tannin-containing genotypes had signifrcantly taller seedling than
genotypes without tannin (Pascual et al., 1990). In their study, Chavan et al., (1981) used
two unrelated lines with high or low tannin content, while in our present study we have a
better control of the genetic background, as progenies of a single cross of two lines
differing in tannin concentration were used. These conflicting findings may then result
from difference in control of the genetic backgrounds.
Seedling vigor scores, germination at 22” C and emergence were significantly
associated with pigments, flavan-4-01s and total phenol content. Seedling height up to 2
weeks afler planting and growth rate 1 (rate of seedling height increase between 7 and 14

---

days afier planting) were significantly correlated with flavan-4-01s and total phenols (Table
3.5). Phenolic compounds have been shown to counteract the inhibitory effect of abcissic
acid (AEL4) on Amaranthus caudatus seedling growth. Compounds with a flavan nucleus
showed a higher activity at low concentration (Ray et al., 1980). Tt was latter shown that
ABA inhibits wheat germination and seedling growth by blocking amylase activity, thus
çhecking the availability of mobilizable carbohydrates essential for these processes
(Sharma et al., 1986). By counteracting the inhibitory effect of ABA, phenoiic
compounds would then increase the availability of mobilizable carbohydrates for
germination and seedling growth. However high concentrations of phenolic compounds
were shown to reduce radish (Raphanus sativus L.) seed germination and corn shoot
growth (Storm 1982). These experiments demonstrating the inhibitory effects of phenolic
çompounds on germination and seedling growth were in vitro studies. If natural inhibitors
are isolated and applied exogenously, they could reach metabolic centers other than those
which they attack in the native state (Kefeli et al., 1971). Thus exogeneous application of
phenolic compounds on sorghum germination and seedlings growth medium as mentioned
by Chukwura et (1982) may not reflect the effect of the endogeneous compounds to the
seeds.
Kernel weight was negatively associated with pigments, flavan-4-01s tannin and
total phenols. These correlations may result from association of these traits in the parental
line Shanqui red which has smaller seeds and higher concentrations of these products than
SRN39.
Only total phenols were significantly correlated with grain yield , while plant
height was not associated with these phenolic compounds (Table 3.6). It was reported in

---

91
chickpea (Citer arietinum) that brown, high tannin lines outyielded white, low tannin
lines. The yield advantage of high tannin lines were attributed to a better stand
establishment resulting from resistance to pathogenic fungi at germination and seedling
stage (Knights et al., 1989). Attack of vegetative tissue by pathogenic fungi results in an
increase in the total phenol content of the tissue (Mi&ra et al., 1980). Infection of
sorghum mesocotyls by Hilminthosporium maydis and Colletotrichum graminicola
resulted in rapid accumulation of the deoxyanthocyanidins, apigenidin and luteolinidin
(Nicholson et al., 1987). Because these compounds were found to be fungitoxic and were
formed only in response to inoculation, they were considered to be phytoalexins and their
synthesis a defense response.
Red coleoptile lines tended to be more vigorous than lines with a green one, as
reflected ‘by the significant differences in seedling vigor scores, percent germination at 22”
c, emergence and seedling height (Table 3.7). Lines with a red coleoptile also had
significantly higher pigment and total phenol content in their seeds than green coleoptile
lines (Table 3.8). Doggett (1988) has classified sorghum seedlings in two groups, red and
green. Red coleoptile color in seedlings is controlled by a single gene with the red Rsr
dominant to green rsr. This study observed 66 red and 33 green coleoptile lines in, with x2
= 10.9, where x2.01= 6.63, showing a, discrepancy of the data with the hypothesized genetic
ratio of 50 % red and 50 % green lines. Flavan-4-01s and tannin content were not
different between the two groups. No significant differences also existed in plant height
and maturity; however red coleoptile lines yielded significantly higher in 1993, and higher
but not significantly SO in 1994 (Table 3.8).

---

Highly significant genetic and additive variantes and high heritability estimates
were found for pigments, flavan-4-01s tannin and total phenols content in this sorghum
population. High levels of phenolic compounds in the grain were associated with vigorous
seedlings, high percent germination at 22” C, emergence, and taller seedhngs. Tannins
have been shown to have antinutritional effects; decreased weight gain in poultry and rats
have been demonstrated (Price et al., 1980). The non significant association of total
phenols with tannin content indicate that there are possibilities to balance the
concentration of these compounds with high seedling vigor through genetic improvement.

---

96
Table 3.1. Expected mean squares, and analyses of variante of phenolic traits.
Sources DF
M S
M S
M S
M S
~-
Pigments
Flavan-4-01s
Tannin
Total phenols
Reps
1
0.001033
0.0000703
0.003421
0.853535
Entries
98
0.023657””
0.0087693
0.013449””
195.536690””
P vs L
1
0.006952”
0.000265”
0.003642””
28.266270**
Retween 1
0.160000””
0.035156””
0.038809”” 676.000000**
P
Among L
9 6
0.02241105
0.008558**
0.013288””
192.274162**
Error
9 8
0 . 0 0 1 3 9 8
0 . 0 0 0 0 9 2
0 . 0 0 1 6 3 6
1.537209
-
-

---

97
c
Table 3 2. Phenolic compounds content of parental and recombinant inbred
lines.
Parental
RI lines
sRN39
SQR
Trait
mean
mean
mean
Range
Pigments
0.0185
0.4185
0.176 3tO.0267
0.001 U-O.4645
”
Flavan-4-01s
0
0.1875
0.068 *0.0067
0.000 - 0.3215
Tannins
0
0.197
0.068 ho.0289
0.000 - 0.3355
phenols
4
3 0
14.314 AO. 0.000 - 47.500

---

Table 3.4. Genetic correlation coefficients and theirs standard errors (in parenthesis)
._
between phenolic compounds.
Trait
Pigments Flavan-4-01s
Tannins
Total phenols
Pigments
0.4655”” (.0829) 0.4218”” (0929) 0.5092”” (.0785)
Flavan-4-01s
0.6154** (0705) 0.5326* (.0730)
Tannins
0.9215”” (.0248)

---

100
Table 3.5. Genetic correlation coefficients and theirs standard errors (in parenthesis) of
phenolic compounds and seedling vigor traits.
Traits
Pigments
Flavan-4-01s
Tannins
Total phenols
Visual scores
-.283* (. 1095)
-.22* (. 1095)
-.241* (.116)
-.331** (.102)
1 OO-seed weight
-.401** (.093)
-.212* (. 1008)
-. 198 (. 108)
-. 143 (.103)
Emergence
,288” (. 1017)
.2189* (1025)
.286* (. 106)
,262” (. 100)
Shoot height 1
-1851 (.1186)
.356** (.lOSS)
.252* (. 121)
.253* (.lll>
Shoot height II
.1394 (.1195)
.3054* (. 1086)
.290* (. 119)
.309** (. 108)
Shoot height III
.2004 (. 1276)
.243 (.1214)
.179 (.134)
.235 (. 122)
Shoot weight 1
.1579 (.1662)
.2678 (.1582)
.285 (. 172)
.446* (. 154)
Shoot weight II
.2137 (.1489)
.1625 (.145)
.156 (.157)
.239 (. 143)
Growth rate 1
.1138 (.1385)
.2798* (. 1284)
.329* (. 138)
.361* (. 124)
Growth rate II
.2849 (.4453)
-.2556 (.4168)
-.505 (.613)
-.337 (.465)
Growth rate III
.2492 (.1812)
.1168 (.1755)
.0967 (. 192)
“144 (.175)
Germin. 112” C
.0344 (. 1272)
-.0877 (. 1225)
-.:314* (. 124)
-. 147 (. 121)
Germin. 22” C
.327** (.1152)
.3524**
.463**
,463”” (.lOZ)
(. 1099)
(.112)

---

101
Table 3.6. Genetic correlation coefficients and theirs standard errors (in parenthesis) of
phenolic compounds and field performance traits.
--
Traits
Grain yield
Plant height
Maturity
Pigments
.2161 (.1505)
-.0723 (. 1042)
-.3657** (.0951)
Flavan-4-,ols
.1723 (. 1476)
-.0863 (1013)
-.2719** (099)
Tannins
.3001 (. 1602)
-.0086 (. 1085)
-.2215* (1081)
Total phenols
.3 148” (. 1432)
.0120 (1019)
-.1609 (.1031)

---

Table 3.7. T test analyses and standard errors (in parenthesis) for coleoptile color
differences in seedling vigor traits.
--
Traits
Green coleoptile
Red coleoptile
Prob > T
--
93 scores
4.37 (0.2576)
3.61 (0.1268)
.0096**
94 scores
2.88 (0.3638)
1.77 (. 1194)
.0052**
93-94 mean scores
3.63 (0.2907)
2.69 (.1046)
.0037**
Emergence
64.91 (2.7260)
72.27 (2.03 10)
.03*
Shoot height 1
12.69 (0.3 15 1)
13.26 (0.1957)
.ll”
Shoot height II
26.67(0.7033)
28.16(0.3853)
.04*
Shoot height III
35.92 (0.7259)
37.68 (0.4542)
.03*
Shoot dry weight 1
324.69 (12.1698)
345.79(6.2026)
.1235"
Shoot dry weight II
841.1 (29.5984)
865.7Q7.7916)
.45"
Germination 12” C
59.33 (4.2283)
59.69(2.6723)
.94"
Germination 22” C
69.83 (2.296)
76.95 (1.4746)
.0076**
,I
-.
._ -_
_-^---
~--,--.--

---

Table 3.8. T test analyses and standard errors (in parenthesis) for coleoptile color
differences in phenolic compounds and field performances.
--
Tsaits
Green coleoptile
Red coleoptile
Prob > T
Pigments
.135 (.0191)
.199 (.0123)
.0045**
Flavan-4-01s
.06(.0104)
.07(.0074)
.24 ns
Tannin
0.49 (.0104)
.08 (.Olll)
.0566"
Total phenols
11.57(1.3404)
15.83 (1.3072)
.02*
93 grain yield
3757.15 (179.9963)
4399.27(131.0333)
.0047**
93 plant height
178.41 (6.4478)
184.64(3.2865)
.38 m
93 maturity
85.47(0.8059)
85.25 (0.4677)
.80 m
94 grain yield
4822.94(158.5308)
5096.85 (126.4629)
.19 ns
94 plant height
180.65 (6.5453)
185.39(3.8195)
0.50"
94 maturity
74.29 (0.7207)
73.64(0.3545)
.415"

---

104
CHAPTER FOUR
IDENTIFICATION OF QUANTITATIVE TRAIT
LOCI ASSOCIATED WITH SEEDLING VIGOR AND CORRELATED
CHARACTERS IN A RECOMBINANT INBRED POPULATION OF SORGHUM

---

105
ABSTRACT
Important advantages have been attributed to seedling vigor, resulting in greater
biomass and grain yield in cold and dry environments. Quantitative trait loti (QTLs) for
estimates of seedling vigor in sorghum (Sorghum bicolor L. Moench.) have been identified
using a recombinant inbred population of a cross between SRN39 x Shanqui red. Q T L
analysis was carried out using the regression approach. RAI?D markers located on linkage
groups D and F were found signiflcantly associated with seedling vigor scores in the two
years of test, explaining the high heritability estimate of this trait. Germination at low
temperature (12” C) and germination-emergence at optimum temperatures were mostly
under digerent genetic control, having only one marker in common, which explained 5 and
6?/0 of their variation. Markers explaining most of the variation for germination and
emergenc,e at optimum temperatures are on linkage group F. Markers on linkage group C
explained 75-80 % of the variation for seedling height. However, only one marker (UE3C
178) on linkage group D, with effect on seedling scores, is associated with seedling height.
AI1 the markers with effect on shoot dry weight were associated with seedling height,
explaining their relatively high genetic correlation. The visual scores and shoot dry weight
ha.d no marker in common. The visual scoring system used to estimate seedling vigor was
then effective in integrating germination, emergence, and seedling height. Two linkage
blocks (D, F) accounted for the differences in seedling vigor in this population. The high
heritability estimate (0.97) of adult plant height between 1993 and 1994 is underlined by
their associated markers which are similar. The high heritability estimate of days to

---

100
.-
maturity çould not be explained by common markers in the 2 years. Four markers were
significantly associated with grain yield in both years. Their contribution to the variation
of yield was reduced by 50 % between 1993 and 1994; this reduction may explain the
moderate heritability (0.49) of grain yield. The identification of markers associated with
seedling vigor and field performance should make breeding for the improvement of these
traits more effrcient, by minimizing the amount of the genotype by environment interaction
effect

---

INTRODUCTION
Until recently, polymorphisms have been detected with phenotypic assays of
genotypes (Dudley, 1993). However genetic analysis based on phenotype is a function of
the heritability of the trait where factors such as the environment and quantitative
inheritance ofien confound the expression of a genic trait. Newly developed DNA
markers have advantages in that they do not have these interactions observed in
phenotype-based assays. Various statistical procedures are employed to correlate
quantitative trait loti (QTL) that control the expression of the trait in question. The
simplest approach for detecting QTL is to analyse the data using one marker at a time.
This simple point analysis does not require a complete molecular linkage map. Lander and
Botstein (1989) proposed the method of interval mapping, to take the fùllest advantage of
linkage maps for quantitative studies. By using linked marker analysis, it is possible to
compensate for recombination between the markers and the QTL. However when the
marker density is high (markers < 15 CM apart) point and interval analysis give nearly
identical results (Tanksley, 1993). Because maximum likelihood estimates reduces to least
squares estimate when data are normally distributed (Snedecor and Co&ran, 1989),
Paterson et al. (1988) and Bubeck et al. (1993) showed that both analytical methods gave
virtually the same results in detecting QTL. Recently, Pereira et al. (1995) identified, with
interval mapping and single-factor analysis of variante, the same unlinked genomic
regions in four different linkage groups with significant effects on sorghum plant height.
lnterval mapping placed the most likely QTL position (likelihood peak) in linkage group

---

A, closer to the RFLP marker ISUI 16. In concordance, single factor andysis ofvariance
indicated ISU116 had the strongest association (P < 0.001) with plant height in the same
linkage group. Similar relationships could be observed for three other linkage groups
containing plant height QTLs.
Similar results were obtained when PROC GLM procedure
of SAS and interval analysis of MAPMAKER were employed for the identification of
QTLs associated with rice blast resistance (Wang et al., 1994). Haley and Knott (1992)
Martinez and Curnow (1994) Wright and Movers (1994) used interval mapping by the
regression approach and obtain similar results as Lander and Botstein (1989) did with the
log-likelihood analysis. However the efficiency of flanking-marker methods decreases as
the number of incompletely-genotyped individuals increases. For this reason Kearsey and
Hyne (1994) proposed a simple “marker-regression” approach. This method produced
estimates of QTL locations and effects comparable to inter-val mapping-approaches based
on log-likelihood or multiple regression. Lark et al., (1994) identified QTLs associated
with maturity and seed oil / protein in a recombinant inbred population of soybean with the
regression analysis, which were the same as those obtained with the distributional
extremes method .
Important advantages have been attributed to seedling vigor, resulting in greater
biomass and grain yield in cold and dry environments (Ludlow et al., 1990). McConnell et
al. (1979) ,attributed the lack of cor-relation between laboratory and field results for cold
emergence and seedling vigor to the mild spring weather during the two years of
evaluation. This lack of cor-relation between laboratory and field results is not surprising
Plant breeders routinely find that genotypes which perform well in one environment are

---

109
not as well suited to another environment. Highly significant interactions of years by
genotypes have been observed for seedling vigor visual scores, grain yield, plant height
and days to maturity as shown in chapters 1 and II of this study. TO minimize the amount
of this genotype by environment interaction effect in future efforts to improve these traits,
we have as the objective of this study the identification of quantitative trait loti associated
with seedling vigor, plant height, maturity and grain yield.

---

110
MATERIALS AND METHODS
Recombinant inbred (RI) lines of a cross between “SRN39’ and ‘Shanqui red’
(SQR) were obtained through the Single Seed Descent method of plant breeding. These
two parents differ in seedling vigor with Shanqui red being the most vigarous. An
experiment employing a randomized complete block design with five replications of 100
RI lines was conducted in 1993 and repeated in 1994 at the Purdue University Agronomy
Research Center in West Lafayette, Indiana. The experimental materials and designs are
the same as described in chapter one. The entries were visually scored for seedling vigor
on a scale of 1 (most vigorous) to 9 (least vigorous). The number of days fiom planting
until anthers had extruded half way down the panicles of at least 50 % of plants in a row
was used as an estimate of maturity. Plant height was measured as centimeters from the
soi1 surface to the top of the head after maturity. Grain yield (kg / ha) was estimated from
tenter 3 m of a plot in 1993, and in 1994 from entire 5 m plot.
The laboratory and the greenhouse experiments conducted, respectively, in an
incubator during February-March, and on a sand bench during June-July 1995. Both had
the same entries, design and procedures for data collections as described in chapter 1.
DNA Preparation
DNA was isolated from Shanqui red, SRN39 and 93 of their recombinant inbred
line progenies. A modification of the CTAB isolation protocol (Bernatsky and Tanksley,
1986) was used. Fifty seeds were germinated at room temperature for five days in moist
*-
.---mm-
-!lmaœœa--
111-

---

111
pa.per towel. Coleoptiles of approximately 3 cm in length from 10-20 seeds of each line
were harvested and placed in sterilized 1.5 ml Eppendorf tube. These samples were frozen
in liquid nitrogen and crushed without thawing, with a metal rod cooled in liquid nitrogen.
0.8 ml CTAB buffer was added to each tube and the tissue samples were homogenized in
the buffer solution using an Eppendorf pellet pestle. Samples were incubated at 550 C for
1 hour an.d then mixed with 0.3 ml phenol / chloroform / isoamyl alcohol(25:24: 1). The
supernatant was recovered by centrifugation and the DNA precipitated using 0.750 ml of
ice cold isopropanol. The DNA pellet was recovered by centrifugation and washed twice
with 70 % ethanol, allowed to air dry, and resuspended in 0.5 ml TE. Aliquots of each
DNA samples were r-un on an agarose gel with known amounts of undigested h DNA and
stained with ethidium bromide. Comparison with the h DNA standards allowed sorghum
DNAs concentration to be estimated. Working stock solutions were then prepared at a
concentration of 5 yg / ml.
RAPD Reactions
Amplification reactions were performed in 25 ~1 volumes containing Taq buffer
(50 mM KCl, 1OmM Tris-HC1 at pH 8.8, 1.5 mM MgCl, 0.1 % Triton x lOO), 0.2 mM
dNTPs, 0.2 mM primer, one unit of taq DNA polymerase,, and 25ng of sorghum genomic
DNA. The reaction mixtures were overlayed with one drop of minera1 oil. Amplification
were performed in an Ericomp Twin Block System Easy Cycler, programmed for 45
cycles of two minutes denaturation at 94” c, one minute annealing at 36’ c and two
minutes elongation at 72” c. Amplification products were analysed by electrophoresis

---

112
through 1.6 % agarose gels for 2.5 hours at 130 volts in TAC buffer (Tris / sodium acetate
/ EDTA). Gels were stained in 1 pg / ml ethidium bromide solutions for 2-4 hours. A
total of 72 primers previously showing polymorphisms between the two parental lines
(Weerasuriya, 1995) were tested. RAPD loti were named OPX#, or UBC#, where OP
and UBC are the primer from Operon and the University of British Columbia, and X and
ff are the primer kit letter and number respectively. When more than one band was scored
for a particular primer, a letter was added at the end of the primer number to identifl each
locus.
Data Analysis
The recombinant inbred lines were scored for RAPD bands that were polymorphie
between the two parents. Shanqui red and SRN39 were scored as 0 and 1 respectively.
RI lines with the Shanqui red or SRN39 allele of a polymorphie RAPD amplification were
scored alternatively as 0 or 1.
An Apple Macintosh version of MAPMAKEZR II was used for linkage analysis.
Linkage groups were constructed with a LOD score > 8.0. A combination of Two point /
Three point and Multipoint / Try commands was used to order the markers within linkage
group with a LOD score > 3. The Kosambi mapping fhnction (Kosambi, 1944) was used
to convert recombination fractions to map distances.
Simple regression analysis was used to identify quantitative trait loti associated
with seedling vigor, yield, plant height, and maturity.

---

113
RESULTS AND DISCUSSION
Fifty one of the 72 primers tested were polymorphie yielding 89 RAPD markers.
Seventy two markers were included in 11 linkage groups, while the map location of 59 of
these have been determined, and 17 markers remained unlinked (Fig. 4.1). The estimated
map size was 925.81 CM, with an average interval between adjacent loci of 19.7 CM. The
maximum distance between any two adjacent markers in this map was 42.9 CM on linkage
group A. Although the locus order and relative genetic distances vary, most of the
linkage groups identified could be integrated to those previously established in another
RAPD map (Weerasuriya, 1995). Tentative correspondences cari be made between the
present linkage groups; D, B and H, C, A, E, 1 respectively with the groups D, P, H, K, Q,
G of the previous map based on marker identification.
Seedling Traits
Several plant characteristics have been associated with seedling vigor . These
include percent germination, percent emergence, seedling height, and dry weight. Visual
scoring has been found to be a convenient and efficient method for estimating seedling
vigor in a breeding program where large numbers of genotypes are evaluated. The
relationships between visual scores and different plant characteristics that estimate
seedling vigor have traditionally been studied with cor-relation techniques (Maiti et al.,
198 1). Recently developed molecular techniques (RAPDs and RFLPs) offer new
possibilities for the genetic study of seedling vigor and methods of estimating the

---

114
phenotype. Regression analysis in this study identified 9 RAPD markers associated with
seedling vigor scores in 1993. These are located in two linkage groups; D and F which
respectively account for 41.21 and 25.78 % of the phenotypic variation (Fig. 4.2). The
same markers were again found significantly associated with vigor scores in 1994 except
only one (OPL9v) in linkage group F. Linkage group D accounted for most of the
variation for seedling vigor with 73.23 % of the total (Table 4.1). This similarity of
markers controlling seedling vigor scores between 1993 and 1994 explain the high
heritability estimate (0.78) of this trait.
Four markers were identified as being significantly associated with germination at
22’ c. The marker OPA16 was the most strongly associated with this trait at the 0.000 1
probability level, and accounted for ‘18.93 % of the variation for germination at normal
temperature. However this marker was not included in any linkage group in this study,
but was mapped previously in the 1 group (Weerasuriya, 1995). Markers on linkage
groups A (OPJlx) and F (OPD16) accounted for 6.07 and 7.56 % of the variation. A
second unlinked marker (OPC7y) was also associated with germination (Table 4.2).
Percent emergence was associated with the unlinked marker OPA16, and with
OPD16 and OPL9v on the linkage group F (Fig. 4.3). The latter two accounted for 16.24
% of the variation for emergence. These three markers with a fourth on linkage group A
(OPJlx) explained 32.35 % of the variation for percent emergence of sorghum seedlings
(Table 4.2). Thus percent germination at 22” c and emergence had most of their
associated markers in common, explainig their highly significant genetic correlation.

---

115
Four markers on linkage group D were significantly associated with germination at
low temperature (12’ c), and accounted for 28.73 % of the variation. Two additional
markers, OPA19 and OPJlx on linkage groups C and A respectively were also associated
with this trait. These 6 markers explained 38.98 % of the variation for germination at low
temperatures. Germination at low temperatures and germination-emergence at normal
temperatures, had thus only one marker (OPJlx) in common with effect on both of them,
on linkage group A and explaining 5 and 6 % of their variation. This means that
germination at low and at normal temperatures are mostly under different genetic control,
explaining the poor selection results obtained, in studies for the improvement of
germination at low temperatures when the environmental conditions in the field are not
appropriate (McConnell et al., 1979). A breeding program for improved germination at
low and normal temperatures should involve a shuttle testing under the two conditions.
The marker OPD16 on linkage group F mostly associated with seedling vigor scores
(P
<: O.OOOl> was also found as having significant effect on germination and emergence at
normal temperatures. A second marker on linkage group F, OPL9v, was also associated
with seedling vigor scores and percent emergence. Two markers on linkage group D
(OPGlOv and OPIlx) with effect on seedling scores, had influences on germination at low
temperatures.
Markers on linkage group C (Fig. 4.4), significantly associated with seedling
height, explained most of the variation (74.91, 79.3, and 39.46 % respectively afier 1, 2,
3 weeks after planting). Additional important linkages groups are A, E and D (Table 4.3).
An unlinked locus (OPKI 8) was also significantly associated with seedling height. This

---

116
marker together with the present group C were previously mapped in linkage group H
(Weerasuriya, 1995). Only one of the marker (UBC 178) on linkage group D, with effect
on. seedling scores, is associated with seedling height. The significant genetic correlation
between germination and emergence at normal temperatures with seedling height and dry
weight could not be explained with any common marker. Although the genetic correlation
between germination at low temperatures and seedling height was not significant, the
markers OPA19 and OPGlOv were associated with both traits.
Seedling dry weight 2 weeks after planting was significantly associated with five
markers on linkage group C (Fig. 4.4). These markers accounted for 32.33 % of the
variation for this trait. In addition, two unlinked markers; OPJOv and OPK18 previously
mapped on linkage group F and H, were also associated with seedling dry weight and
accounted for 6.1 and 5.49 % of the variation respectively. Only OPJ6v and OPKI 7 on
linkage group C were associated with seedling dry weight 3 weeks alter planting and
ac.counted for 5.05 and 5.71 % of the variation (Table 4.4). Al1 the markers with effect on
seedling dry weight were then assoçiated with seedling height explaining their relatively
high genetic correlation. It cari then be stated again that efficient selection for seedling
height would improve seedling dry weight. Progress in breeding for improved seedling
dry weiglt is expected to be slow as suggested by a low to moderate heritability estimate.
There was no marker in common between seedling dry weight and the visual scores.
The visual scoring system used to estimate seedling vigor was then effective in
integrating germination, emergence, and seedling height. This is consistent with the
conclusion that seedling-visual scores are controlled by 2-3 genes. It appears now more

---

117
appropiate to state that two linkage blocks account for the differences in seedling-vigor
scores in this population. The markers associated with germination, emergence and
seedling height explained about 25 % of the variation of the scores. However, the scores
did not account for the differences in seedling dry weight among lines. Most of the
markers on linkage group D, with effect on the visual scores, were not associated with any
of the seedling traits measured.
Plant characteristics which have been associated with seedling vigor (Lopez-
Castaneda et al., 1995) and with possible relevance for the visual scores include, rate o.f
germination, coleoptile length, and rate of leaf production. Perhaps these traits might
explain the additional variation in seedling scores.
Adult Plant Traits
Most of the variation for plant height (Table 4.5) in each year could be explained
by markers on linkage group H (90.4 and 88.78 % respectively). Markers on linkage
groups A, B and G were also significantly associated with plant height (Fig. 4.5). Seedling
height one week afier planting, had one marker in common with adult plant height. This
marker (OPEl) accounted for about 5 % of the variation in the two traits, explaining their
low and non-significant genetic correlation. However the high heritability estimate
(0.9754) of adult plant height is underlined by their associated FUPD markers which are
similar. Three of these, (OPC2Ox; OPE1 and OPL3v) have been previously found with
effect on plant height, through interval-mapping and maximum-likelihood method
(Weerasuriya, 1995). These findings also suggest that the two methods of analysis give

---

118
similar results. The marker (UBC122) previously providing the fourth linked marker was
not tested in this study. Pereira et a1 (1995) using RFLP, identified four unlinked genomic
regions with significant effects on sorghum plant height. A standardization of
methodology may help to establish if these four unlinked genomic regions are the same as
the ones identified in the present study.
In the 1993 data, five markers on Iinkage group D explained 32.52 % of the
variation for days to maturity (Fig 4.5). Two additional markers had an effect on this trait
in both ‘1993 and 1994. These are OPL3x on linkage group H with 5.58 and 5.02 %, and
OPE1 8v with 7.8 1 and 5.76 % of the variation of days to maturity in 1993, and 1994
respectively (Table 4.6). Although the heritability estimate of days to maturity was
relatively high (0.8890) the linkage group D with no effect on the second year accounted
for most of the variation (32.52 %) in the first one. However, they shared two associated
markers, accounting for approximately 1 O- 12 % of their variation.
Grain yield was significantly associated with four markers in linkage group H,
which accounted for 5 1.53 % of the variation in 1993 (Fig 4.5). Two additional markers;
OPA1 6 and OPJOv explained 8.2, and 9.71 % respectively of the variation for yield (Table
4.7). Two of these markers on linkage group H were again significantly associated with
grain yield in 1994, but they explained much less of the variation (11.92 OA). Markers
OPD16 on linkage group F and the unlinked OPJOv were also associated with grain yield
in both years and accounted for 11.32, 5.20 and 9.7lin 1993 and 1994, and 6.65 % in
1993 and 1994 respectively. Marker OPEl4v on linkage group A was most closely
associated with grain yield in the second year (lO”h of variation; p < 0.0025). Markers

---

OPC7v; OPCZOx; OPD16 and OPDJ6v were significantly associated with grain yield in
both years. Their contribution to the variation of yield was reduced by 50 % from 1993
to1994 i.e., in 1994 compare to 1993 from 51.53 % to 23.77 %. This reduction may
explain the moderate heritability estimate of grain yield (0.4975). Days to maturity and
grain yield of 1993 had one marker (OPL3x) in common which accounted for about 5-9 %
of their variation , explaining the non-significant genetic correlation. The marker OPL3x
was also an important contributor (23 OA) to the variation of plant height; but the genetic
correlation of this character and days to maturity was negative and non significant. This
situation could be explained by the fact that the ta11 parent (Shanqui red) was also earlier.
Markers on linkage group C were found significantly associated with flavan-4-ols,
pigment, tannin, and total phenols (Table 4.8). These common markers may explain the
Sign&ant genetic correlations between the phenolic compounds. Three markers, OPA 19,
OPB16x, and OPK17 accounted for 19.95, 26.51, 30.77 % respectivly of flavan-4-ols,
tannin, and total phenol variation. Markers OPC7x and OPE8y on linkage group A and B,
and mark.er OPK17 were associated with pigment content,. Flavan-4-ols, tannin, and total
phenols have the same linked markers on group C as seedling height, explaining their
significant genetic correlations (Fig 4.6). Although these markers were also associated
with seedling dry weight, only total phenols were significantly correlated with this trait.
The significant correlation of germination at 22” C and the phenolic compounds could net
be explained by common markers. But the marker OPA19 which is important for
phenolics was also associated with germination at 12” C.

---

120
The basis of heritability estimates and genetic cor-relation coefficients among traits
have been explained with the identification of DNA markers. Markers associated with
seedling vigor have been identified. This should make breeding for improved germination
and enhanced seedling growth at low and normal temperatures more efficient, by
mmimizing the amount of the genotype by environment interaction effect on the
expression of genotypes.
-.
..,
.--.

.“w..z.““IU

--m-mw-*I--
U’

---

1 2 1
Table 4.1. Genetic markers that significantly cosegregate with seedling vigor scores
in (SRN39 x SQR) recombinant inbred population.
1993 scores
1994 scores
Marker
Prob > F
R-square
Prob > F
R-square
linkage group
O P A 2
0 . 0 1 5 6
0.0684
0.0016
0.1146
D
OPC2Oy
0 . 0 3 2 6
0.0539
0.0030
0.1025
D
OPD8y
0 . 0 1 2 7
0.0726
0.0015
0 . 1 1 6 7
D
OPD16
0.0001
0.1617
0.0165
0.0681
F
OPGl Ov
0.0333
0.0483
0.0008
0.1291
D
OPIlx
0 . 0 2 8 6
0 . 0 5 6 4
0.0041
0.0964
D
OPJ12x
0 . 0 3 8 6
0 . 0 5 0 6
0.0032
0 . 1 0 0 9
D
OPLBV
0 . 0 0 3 9
0.0961
ns
F
UIC178
0 . 0 2 1 7
0 . 0 6 1 9
0.0135
0.0721
D
_I
_
s --
- “- ,Tw..,-
--“IIILp--I
-i

---

122
Table 4.2. RAPD markers associated with germination and emergence in (SRN39 x
SQR) RI population.
Germination 12”~
Germination 22”~
Emergence
Marker
Prob > F R-square Prob > F R-square Prob > F R-square group
-
-
OPA16 ns
-
0.0001
0.1893
0.0043
0 . 0 9 1 2
unlked
OPA19 0.0437
0.0476 ns
-
ns
C
OPC7y ns
-
0 . 0 0 7 6
0.0799 ns
-
unlked
OPC2Oy 0 . 0 0 3 2
0.0989 ns
-
ns
D
OPD16 ns
-
0.0095
0.0756
0.003
0.0998 F
OPGlOv 0.0365
0.0510 ns
-
ns
D
OPIlv
0 . 0 0 5 0
0.0900 ns
-
ns
D
OPllx
0 . 0 4 4 0
0.0474 ns
-
ns
D
OPJlx
0 . 0 2 9 8
0 . 0 5 4 9
0.0118
0 . 0 6 0 7
0 . 0 1 3 9
0.0699 A
OPL9v ns
-
ns
0.0191
0.0626 F
-
-

---

123
?‘a,ble 4.3. RAPD markers associated with seedling height in (SRN39 x SQR) R.I Pop.
-
-
height 1
height 2
height 3
Marker
P r o b > R-
P r o b > R-
Prob > R-
group
F
square F
square F
square
-
-
O P A 1 9
0.002 0.1289
0.0001 0.1468
0.0027 0.1070
c
OPB16x
0.0001 0.1870
0.0001 0.1756
0.0022 0.1117
c
OPC7x
0.0429 0.0391
ns
-
ns
A
OPClOv
0.0018 0.0904
0.0021 0.0759
ns
-
C
OPE1
0.0214 0.0502
ns
-
ns
A
OPGlOv ns
-
ns
0.0404 0.0515
D
OPJlv
ns
-
0.0180 0.0459
0.0229 0,063 1
unlinked
OPG6v
ns
-
0.0423 0.0342
ns
-
unlinked
OPJ12.v ns
-
ns
0.0400 0.0517
D
OPJ17v
0.0460 0.0381
0.0267 0.0405
ns
-
E
O P K 8
0.0011 0.0976
0.0002 0.1061
0.0218 0.0640
C
O P K 1 7
0.0001 0.1622
0.0001 0.1858
0.0021 0.1119
c
O P K 1 8
0.0034 0.0795
0.0010 0.0862
0.0338 0.055 1
unlinked
OPL9x
0.0305 0.0445
ns
-
ns
G
OPM2v
0.0028 0.0829
0.0035 0.0688
ns
-
C
OPM2x
ns
-
ns
0.0202 0.0656
m
DBC178 ns
-
0.0429 0.0340
0 . 1 1 6 0 . 0 7 7 1
D

---

124
.^
. .
‘Table 4.4. RAPD markers significantly associated with seedling dry weight in
@RN39 x SQR) RI population.
Dry weight 1
Dry weight 2
Marker
Prob > F
R-#square
Prob > F
R-square
linkage
group
O P A 1 9
0 . 0 4 2 6
0.0492 ns
C
OPB16x
0 . 0 0 5 7
0.0894 ns
C
I
OPC7x
0 . 0 0 4 7
0.0936 ns
A
OPClOv
0 . 0 4 8 9
0.0465 ns
c
OP.TGv
0.0235
0.0610
0.0345
0 . 0 5 0 9
unlinked
OPR8
0.0465
0.0475
ns
-
c
O P K 1 7
0 . 0 0 5 4
0.0907
0.0250
0.0571
c
.C
OPK18
0 . 0 3 2 0
0.0549 ns
unlinked

---

125
Table 4.5. RAPD markers significantly associated with plant height in (SRN39 x SQR)
RI population. Parenthesis indicated that marker location was not determined.
93-plant height
94-plant height
‘Marker
Prob > F
R-square
Prob > F
R-square
linkage group
-~
O P A 2 0
0 . 0 0 0 2
0 . 0 8 9 4
0.0001
0 . 0 9 4 4
H
OPBlO
ns
0.0364
0 . 0 2 3 6
G
OPC7v
0.0001
0.1408
0.0001
0.1548
H
OPC2Ox
0.0001
0.1611
0.0001
0.1701
H
OPE1
0.0175
0.377
0.0153
0.03 14
(4
OPE14x
0.0283
0 . 0 3 2 2
0.0321
0 . 0 2 4 7
(W
OPJI 8x
0.0498
0 . 0 2 5 9
0.0137
0 . 0 3 2 4
B
OPL3v
0.0015
0.0679
0.0020
0.0513
H
OPL3x
0.0001
0.2396
0.0001
0 . 2 0 5 6
H
OPL3y
0.0001
0 . 1 2 5 6
0.0001
0 . 1 3 8 9
H
OPL19
0.0005
0.0796
0 . 0 0 0 2
0 . 0 7 2 7
H
.-
,“*
_I
*a
. I
- -
. I “..w-
-LIM
-*m--m
-1

---

126
Table 4.6. RAPD markers significantly associated with maturity in (SRN39 x SQR Rl)
. .
population.
93 -plant maturity
94-plant maturity
Marker
Prob > F
R-square
Prob > F
R-square
linkage group
-
-
OPD3x
ns
0.0228
0.0575
A
OPDSy
0 . 0 0 7 4
0.0763
ns
D
OPE8y
.00121
0.0672
ns
B
OPE18v
0.0064
0.0781
0.0228
0.0576
unlinked
OPllx
0.0135
0.0652
ns
D
OPJ12v
0 . 0 2 2 0
0.0558
ns
D
OPJ12x
0 . 0 0 8 9
0.0727
ns
D
OPL3x
0.0192
0.0581
0.0338
0 . 0 5 0 2
H
DBC178
0 . 0 2 3 4
0.0552
ns
D

---

127
Table 4.7. RAPD markers significantly associated with grain yield in (SRN39 x SQR)
population.
-
93 grain yield
94 grain yield
Marker
Prob > F
R-square
Prob > F
R-square
linkage group
-
-
O P A 1 6
0 . 0 0 6 2
0.0820 ns
unlinked
O P A 2 0
0 . 0 0 0 9
0.1186 ns
H
OPC7v
0.0001
0.1665
0.0265
0.0553 H
OPC2Ox
0.0003
0.1368
0.0168
0.0639
H
O P D 1 6
0 . 0 0 1 2
0 . 1 1 3 2
0 . 0 3 1 6
0 . 0 5 2 0
F
OPE14v
ns
-
0.0025
0 . 1 0 0 0
A
OPJOV
0.0028
0.0971
0.0147
0.0665
unlinked
OPL3x
0 . 0 0 3 4
0.0934
ns
H

---

128
Table 4.8. RAPD markers significantly associated with phenolic compounds.
Flavan-4-01s
Pigments
Market
Prob > F
R-square
Prob > F
R-square
linkage Group
-OPA19
0 . 0 0 9 7
0.0798
C
OPBlGx
0 . 0 3 2 4
0.0553
c
OPC7x
-
0.0071
0 . 0 8 6 0
A
OPE8y
-
0.0138
0.0725
B
”
OP113
0 . 0 0 5 9
0 . 0 9 0 0
unlinked
OPJ17x
0 . 0 3 0 9
0.0562
unlinked
OPK8
0.005 1
e
0.0928
C
OPK1’7
0 . 0 2 0 6
0 . 0 6 4 4
0.0425
0.0498
C
. .
-.
. . .

---

129
Table 4.8 cent. RAPD markers significantly associated with phenolic compounds
Tannin
Total phenol
Marker
Prob > F
R-square
Prob > F
R-square
Linkage group
OPA19
0.0037
0.0993
0.0124
0.047
C
OPB16x
0.0134
0.0731
0.0014
0.1185
c
OPClOv
-
0.0049
0.0935
C
OPK17
0.0051
0.0927
0.0004
0.1422
c
OPM2v
-
0.0030
0.1033
C

---

130
-OPA19
-OPA2
~UBC193y
OPE8z
14.1
16.4
-0pBlêx
-0PBlGv
20.5
36X
-0PE14V
2 0 . 0
405
28.l 3
21.:
.oPEBv
-0PC2Q
7 . 8
- OPLl 1
-OFclov
-0PD8y
6 . 3
-0PGlov
14.’
27.4
32.:
-0PC6
22.41
12.!
-0PGlox
-0PJl8x
-0Pc7x
-0PllV
i
29.3
17.5
-0PJ13
-0PK8
-0PEBy
-0PllX
-0PJI8v
9.t ;
1 7 .
-0PJl2v
3
5.:3
-0PJlx
-0PJl2x
-0PL8x
-0PEBr
9.4
- UBCI 78
27.4
24.
-0PGl8
-0PKl7
6 .
-OP14
H
4.
- OPE&
-’ OPA20
-UBGl25
- IJBCl58v
J
17.
8 . 4
1 . 2
- oPc2ox
-0PD8x
OPL3Z
-0Ph42v
-xPC7v
125
-0PD3x
-. OPL3x
30
20.9
15.;
22
-0PBlO
- OPL3y
EuBc1932
8 . 2
-0PJ6v
uBc193x
-0PD3v
19.: )
F
2 7.8
- OPLl9
CIOX
O P D 1 6
t
1 4 .
7,
c- OPL9v
-ul3c171
-0PL3X
120
4 2
-OPE1
Figure 4.1. A genetic linkage map of sorghum with 59 markers in 11 linkage goups.
Numbers on left of linkage groups represent map distances in centi-Morgans.

---

131
E
-0PB8x
-OPA19
- OPA2
.uBcM3y
-OPE62
-OP01 6x
-0PB16v
-0PEl4V
-0PE8v
-0FC2OY
-0FG8x
. OPLII
-0PCIov
-0PDBy
-0PGlov
-0PC6
-0PGlox
-0PJl8x
-0Pc7x
-0PllV
-0PJ13
-0PK8
-OPE&
-0PllX
-0PJlSV
-0PJl2v
-0PJlx
-0PL8x
-0PEBr
-0PJl2x
- UBCl78
-0PGl8
-0PK17
-OP14
l-i
- OPB8v
- OPA20
-UBCl25
- uEm58v
- oPc2ox
J
-0PD8X
OPL3Z
-0PMïv
--b7v
-0PD3x
- OPL3X
-0PBlO
- OPL3y
uBc193z
-0PJ6v
UBCl93x
-0PD3v
I’
K
F
- OPLi
OPCIOX
OPD16
t
slr OPLSV
-UBC171
-0PLQx
OP120
-OPE1
Figure 4.2. Location of QTLs for seedling vigor scores in SRN39 x SQR RI population.

---

132
C
E
-OPA19
-0PA2
-0PB8x
-UBC193y
-0pEaz
1
-0PB16x
-0PB16v
-0PE14v
-0Fav
-0PC2W
-0PG8x
. oPLi
-0PClOV
-0pDsy
-0PGIov
-OF%%
-0PGlox
-0PJlsX
*
-0Pc7x
-0PllV
-0PJ13
-0PK8
-0PE8y
-0PllX
-0PJ18v
-0PJ12v
-0PL6x
-0PJlx
-0PJl2x
-0PE8f
- UBC178
-0PG18
-0PK17
-0Pl4
H
- OPB8v
-OPA20
-UBG125
- uBc158v
- oPc2ox
J
-0PDax
OPL3.z
-0PM2v
--bC7V
- oPL3x
-0m3x
G
-0PBlO
- 0PL3y
UBCI 932
lJBc190<
-0PDâV
F
K
- WL19
Clox
OPD16
/
OPL9v
-UBC171
-0PL9x
IIl2.0
i
n Gemiinatim at 12 c
Germination at 22 c
Emergence
-OPE1
Figure 4.3. Location of QTLs for germination and emergence in SRN39 x SQR RI pop.

---

1 3 3
-OPA19
-0PB8x
- W C 1 93y
-OPE82
-OPA2
-0PBl6x
- OPE1 6v
-0PE14v
-0PExw
-0Pc2oy
-0PGax
- OPLI 1
-0PCl o v
-0PD8y
-0ffiIOV
-0pG6
-0PJ18x
-0PGlox
-0Pc7x
-0Pllv
-Of’K8
-0PJl3
-0PE8y
-0Pllx
-0PJlBv
-0PJ12v
-0PJlx
-OPE&
-0PJ12x
-0PLBX
- IJBC178
-0PG18
-0PK17
-OP14
- OPmv
- OPA20
- uBc15av
-UBCl25
- oF+c2ox
J
-0PDBx
-0PM2v
XPC7v
OPL32
-0w3x
- - OPL3x
-0PBlO
- oPL3y
uBc1932
-0PJGv
-0PD3v
r uBc193x
F
K
- OPL19
PClox
O P D 1 6
t= OPL9v
-IJBC171
-0PL9X
POP120
Seedling height
-OPE1
Figure 4.4. Location of QTLs for seedling height, dry matter., in SRN39 x SQR RI pop.

---

1 3 4
-UBC193y
-0PE6z
-OPA19
-OPA2
-0PB6x
E
-0PBlGx
-0PB16v
-0PE14v
-0PC2O)J
-0PG6x
- OPLll
-0PClov
-0PDBy
-0PGloV
-0PC6
-0%7x
-0PJ16x
-0PG1ox
-0Pllv
-0PK6
-0PJ13
-0PESy
-0Pllx
-0PJl6v
-0PJ12v
-0PJlx
-OPE&
-0F’JlZx
-0PL6x
- UEIC176
-Offi
-0PK17
‘OP14
- OPBBV
- uBc156v
lsl H OPA20
-UBC125
ofX2ox
.OPLm
-0PM2V
L c7v
-0PL32
-0m3x
oPL3x
OPL3y
-uBc1932
‘OPJ6v
-0PD3v
-uBc193x
OPDIG
OPL19
OPL9v
“UBC171
l-OPL9x
Matufity
Yield
Plant heightt
-OPE1
Figure 4.5. Location of QTLs for maturity, height, yield, in SRN39 x SQR R1 pop.

---

-OPA19
-0PBax
-UBC193y
-OPE82
-OPA2
-0PBlGx
-0PB16v
-C?PEl4v
-0PE8v
-0PC2oy
-0PGax
- OPLI 1
-0PCl o v
-0PD8y
-0PGloV
-0PC6
-0PJl8x
-0PGlox
-0Pf27x
-0PllV
-0PK8
-0PE8y
-0PllX
-0PJlhr
-0PJlx
-0PEBr
-0PJl2x
-UBC178
-0PGl8
-0PK17
-0P14
- OP08v
- OPA20
- uBc1m
- oPc2ox
-0W8x
-0PM2v
Tnx7v
-0PD3x
- OPL3X
-OP610
- oPL3y
-0PJ6v
-0PDJv
K
- WL19
PClox
OPD16
OPL9w
-UBCl71
-0PL9-x
POPIM
e Flavand-ols
Pigments
FlavanA-ols
Q
-OPE1
Total ptienols
Figure 4.6. Location of QTLs associated with phenolic compounds.

---

136
REFERENCES
Abdalla, F. H.,and E. H. Roberts. 1969. The effect of seed storage conditions on the
growth and yield of barley, broad beans, and peas, Ann. Bot. (London)
33:169-184.
Adegbuyi, E., and J. S. Burris. 1989. Effects of seed vigor on trop characters in uniform
and reduced populations of cor-n (.&a mays L.). J. Agron. Crop Sci. 162: 10-20.
Abdullahi, A., and R. L., Vanderlip. 1972. Relationships of vigor tests and seed source
and size to soybean seedling establishment. Agron. J. 64: 143-144.
Acevedo, E., P. 0. Crawfurd, R. B. Austin, and P. Perez-Marco. 1991. Traits associated
with high yield in barley in low-rainfall environments. J. Agric. Sci. (Cambridge)
116:23-36.
Allard, R. W. 1988. Genetic changes associated with the evolution of adaptedness in
cultivated plants and their wild progenitors. J. Hered. 79:225-238.
Anda, A., and L. Pinter. 1994. Sorghum germination and development as influenced by
soi1 temperature and water content. Agron. J. 86:621-624.
Angus, J. F., R. B. Cunningham, M. W., Moncur, and D. H. Mackenzie. 198 1. Phasic
development in field crops. 1. Thermal response in the seedling phase. Field
Crops Res. 3:365-378.
Anon. 1979. Report of a rice cold tolerance workshop. Intern. Rice Res. Inst. Los
Banos, Philipines, ~~112.

---

137
Bacon, R. K., R. P. Cantrell, and J. D. Axtell. 1986. Selection for seedling cold
tolerance in grain sorghum. Crop Sci. 26:900-903.
Barla-Szabo, G., J. Bocsi, B. Dolinka, and M. 0. Diema. 1990. Diallel analysis of seed
vigor in maize. Seed SC. Teçhnol. 18:721-729.
Bennett, M. A.,and L. Waters, Jr. 1984. Influence of seed moisture on lima bean stand
establishment and growth. J. Am. Soc. Hortic. Sci. 109:623-626.
Bennetzen, J. L.,and M. Freeling. 1993. Grasses as a single genetic system: genome
composition, collinearity and compatibility. Trends in Genetics, 9:259-261.
Bernatzky, R., S. D. Tanksley. 1986. Method for detection of single or low copy
sequences in tomato on Southern blots. Plant Mol. Biol. Rep. 4:555-557.
Blum A. 1988. Plant breeding for stress environments. CRC Press, Inc. Boca Raton,
Florida. 223~~.
Brar, G. S., and B. A. Stewart. 1994. Germination under controlled temperature and
field emergence of 13 sorghum cultivars. Crop Sci. 34: 1336-1340.
Brown, A. L. 1986. Breeding corn for earliness. Proc. Am. Corn Sorghum Res. Conf.,
23:59-66.
Bubeck, D. M., M. M. Goodman, W. D. Beavis, and D. Grant. 1993. Quantitative trait
loti controlling resistance to gray leaf spot in maize. Crop Sci. 33:838-847.
Burnham, C. R. 1966. Cytogenetics in plant improvement. Iowa State Press. Ames, IA
Burr, B., F. A. Burr, K. H. Thompson, M. C. Albertsen, and C. W. Stuber. 1988. Gene
mapping with recombinant inbreds in maize. Genetics 118: 5 19-526.

---

13X
Burr, B., F. A. Burr. 199 1. Recombinant inbreds for molecular mapping in maize:
Theoretical and practical considerations. Trends Genet. 7:55-60.
Burris, J. S. 1975. Seedling vigor and its effect on field production of corn Pr-oc. Cern.
Res. Conf 30: 185-193.
Butler, L. G. 1989. Sorghum polyphenols. In Toxicants of plant origin, Volume IV;
phenolics, P. R. Cheeke, Ed., CRC Press, Boca. 95-121.
Camargo,, C. P., and L. E. Vaughan. 1973. Effect of seed vigor performance and yield of
grain sorghum (Sorghum bicolor L. Moench.) Process. Assoc. Off. Anal.
63: 135-147.
Cattivelli,, L.,D Bar-tels. 1989. Cold-induced mRNAs accumulate with different kinetics
in barley coleoptiles. Planta 178: 184-l 88.
Charles-Edwards, D. A. 1982. Physiological determinants of trop growth. Academic
Press Australia, North Ryde, NSW.
Chavan, J. K., S. S. Kadam and D. K. Salunkhe. 1981. Changes in tannin, free amino
acids, reducing sugars, and starch during seed germination of low and high tannin
cultivars of sorghum. J. Food Sci. 46:638-639.
Chittenden, L. M., K. F., Schertz, Y. R. Lin, R. A. Wing, A. H. Paterson. 1994. A
detailed RFLP map of Sorghum bicolor x S. propinquum, suitable for high density
mapping, suggests ancestral duplication of sorghum chromosomes or
chromosomal segments. Theor. Appl. Genet., 87:925-933.
Chukwura, E. N., H. G. Muller. 1982. Effect of tannic acid on a low tannin African
sorghum variety in relation to carbohydrate and amylase. J. Food Sci.

---

139
4’7:1380-1381.
Churcill, G. A., and R. W. Doerge. 1994. Empirical threshold values for quantitative
trait mapping. Genet. 138:963-971.
Condon, .A. G., R. A. Richards, and G. D. Farquhar. 1987. Carbon isotope
discrimination is positively correlated with grain yield and dry matter production
in field grown wheat. Crop Sci. 27:996-1001.
Cooper, P. J. M., P. J. Gregory, J. D. H. Keatings, and S. C. Brown. 1987. Effects of
fertilizer, variety, and location on barley production under rainfed conditions in
northern Syria. II Soi1 water dynamics and trop water use. Field Crops Res.
16:67-84.
Dalvi, R. R. ‘1974. Biochemical and physiological studies on phytotoxicity of selected
pesticides and allergens during seed germination of some food crops. PH.D.
dissertation, Utah State Univ., Logan Utah.
Dogget, H. 1988. Sorghum. Longman Group, UK.
Dudley, J. W., and R. H. Moll. 1969. Interpretation and use of estimates of heritability
and genetic variantes in plant breeding. Crop Sci. 9:257-266.
Dudley, J. W. 1993. Molecular markers in plant improvement: Manipulation of genes
affecting quantitative traits. Crop Sci. 33:660-668.
Egli, D. 13., D. M. Tekrony, and R. R. Wiralaga. 1990. Effect of soybean seed vigor and
size on seedling growth. 5. Seed Technol. 14: 1-12.
Ellis, R. II. 1989. The effects of differences in seed quality resulting fi-om priming or
deterioration on the relative growth rate of onion seedlings. Acta Hort.

---

140
253:203-211.
Ellis. R. II, 1992. Seed and seedling vigor in relation to trop growth and yield. Plant
growth regulation 11:249-255.
Evans, L. E., and K. F. Best. 1977. The effect of soi1 water potential, temperature and
seedling depth on seedling emergence of wheat. Can. J. Plant Sci. 53:43-46.
Everson, E. H., and C. W. Schaller. 1955. The genetics of yield differences associated
with awn barbing in the barley hybrid (Lion x Atlas 10) x Atlas. Agron. J.
47:276-280.
Falconer, D. S. 1989. Introduction to quantitative genetics. Third edition. John Wiley
& Sons, New York, 438 p.
Fisher, R.. A. 1935. The design of experiments, Ed. 3. Oliver and Boyd Ltd, London.
Fisher, R.. A., and J. T. Wood. 1979. Drought resistance in spring wheat cultivars. III.
Yield associations with morpho-physiological traits. Aust. J. Agric. Res.
30:1001-1020.
Funk, C. R., J. C. Anderson, R. W. Atkinson. 1962. Effect of seed source and seed age
on field and laboratory performance of field corn. ‘Crop Sci. 2:318-320.
Gilmour, S. J., R. K. Hajela, M. F. Thomashaw. 1988. Cold acclimation in Arabidopsis
thaliana. Plant Physiol. 871745-750.
Gopinath, A., and H. S. Shetty. 1985. Occurrence and location of Fusarium species in
Indian sorghum seed. Seed Sci. and Technol. 13:521-528.
Gordon, A. G. 1971. The germination resistance test, a new test for measuring
germination quality of cereals. Can. J. Plant Sci. 5 1: 18 l- 183.

---

1 4 1
Gregory, P. J., D. Tennant, and R. R. Belford. 1992. Root and shoot growth, and water
and light use effrciency of barley and wheat crops grown on a shallow duplex soi1
in a mediterranean type environment. Aust. J. Agri. Res. 43 :555-573.
Guy, C. L. 1990. Cold acclimation and freezing stress tolerance: Role of protein
metabolism. Ann. Rev. Plant Physiol. Plant Mol. Biol. 41: 187-223.
Haley, C. S., and S. A. Knott. 1992. A simple regression method for mapping
quantitative trait loti in line crosses using flanking markers. Hereditas
69:3 15-324.
Hahn,D. H., L. W. Rooney, C. F. Earp. 1984. Tannins and Phenols of sorghum. Cereals
Food World, 29:776-779.
Hahn,D. H., J. M. Faubion, and L. W. Rooney. 1983. Sorghum phenolic acids, Their
high performance liquid chromatography seperation and their relation to fungal
disease. Cereal Chemistry 60: 255-259.
Hahn,D. H., and L. W. Rooney. 1986. Effect of genotype on tannins and phenols of
sorghum. Cereal Chemistry 63 :4-8.
Hanson, W. D. 1959. The breakup of initial linkage blocks under selected mating
systems. Genetics 44:857-868.
Harrison, B. J. 1966. Seed deterioration in relation to seed storage conditions and its
influence upon seed germination, chromosomal damage and plant performance. J.
Natl. Inst. Agric. Bot. (UK) 10:644-663.
Hoshikawa, K. 1969. Underground organs of the seedlings and the systematics of
gramineae. Bot, Gaz. 130:192-203.
.~.
-_._--. -
--.-.-.-. -..-----

---

112
Hulbert, S. H., T. E. Richter, J. D. Axtell, and J. L. Bennetzen. !990. Genetic mapping
and characterization of sorghum and related crops by means of maize DNA
probes. Proc. Natl. Acad. Sci. 87:4251-4255.
Ibrahim, 0. E., W. E. Nyquist, J. D. Axtell. 1985. Quantitative Inheritance and
correlations of agronomie and grain quality traits in shorghum. Crop Science
25:649-654.
Jones, D. B., M. L. Peterson. 1976. Rice seedling vigor at sub-optimal temperatures.
Crop Sci. 16:102-105.
Ju.mbunathan, R., and E. T. Mertz. 1973. Relationship between tannin level, rat growth,
and distribution of protein in sorghum. J. Agri. Food Chem. 22: 1156-
Jumbunathan, R., M. S. Kherdekar, YR. Bandyopadyay. 1990. Flavan-4-01s concentration
in mold susceptible and mold. resistance sorghum at different stages of grain
development. J. Agric. Food Chem. 38:545-548.
Kalburtji, K. L., J. A. Mosjidis. 1993. Effects of sericea lespedeza residues on cool-
season grasses. J. Range Manag. 64:3 15-3 19.
Kambal, A. E., and E. C. Bate-Smith. 1976. A genetic and biochemical study on
pericarp pigments in a cross between two cultivars of sorghum, Sorghum bicolur.
Heredity 37:413-416.
Kearsey, H. J., V. Hyne. 1994. QTL analysis; a simple “marker-regressi’on approach.
Theor. Appl. Genet. 89:698-702.
Kefeli, V. J., C. H. Kadyrov. 1971. Natural growth inhibitors, theirs chemical and
physiological properties. Ann. Rev. Plant Physiol. 22: 185- 196.

---

143
Keim, K., R., and C. 0. Gardner. 1984. Genetic variation for cold tolerance in selected
and
unselected maize populations. Field Crop Res. 8: 143- 15 1.
Khah, E. M., E. H. Robe&. 1986. Effects of laboratory germination, soi1 temperature
and moisture content on the emergence of spring wheat. J. Agric. Sci., Camb.
107:431-438.
Khah, E. M., E. H. Roberts, and R. H. Ellis. 1989. Effects of seed aging on growth and
yield of spring Wheat at different plant-population densities. Field Crop Res.
20:175-190.
Knights, :E. J., R. J. Mailer. !989. Association of seed type and color with establishment,
yield and seed quality in chicpea (Citer aerietimm). J. Agric. Sci. Cambridge.
113:325-330.
Kosambi, D. D. 1944. The estimation of map distances from recombination values.
Annals ofEugenics 12:172-175.
Kruglyak, L., Lander, E. S. 1995. A nonparametric approach for mapping quantitative
trait loti. Genet. 139: 1421-1428.
Lafond, G. P. ,and R. J. Baker. 1988. Effects of genotype and seed size on speed of
emergence and seedling vigor in nine spring wheat cultivars. Crop Sci.
26:341-345.
Lander, 13. S., R. V. Kesseli, B. Farrara, and R. W. Michelmore. 1987. A genetic map of
lettuce (Lactuca sativa L. ) with restriction fragment length polymorphism,
isozyme, disease resistance, and morphological markers. Genet& 116: 3 3 l-33 7.

---

144
Lander, E. S., and D. Botstein. 1989. Mapping Mendelian factors underlying
quantitative traits using RFLP linkage maps. Genetics 12 1: 185 199.
Law, c. N’, 1967, The location of factors controlling a number of quantitative characters
in ,wheat. Genetics 56:445-461.
Lawlor, D. J., E. T. Kanemasu, W. C. Albrecht III, and D. E. Jonshon. 1990. Seed
production environment influence on the base temperature for growth of sorghum
genotypes. Agron. J. 82:643-647.
Lebowitz, R. J., M. Soller, and J. S. Beckmann. 1987. Trait-based analyses for the
detection of linkage between marker loti and quantitative trait loti in crosses
between inbred lines. Theor. Appl. Genet. 73:556-562.
Liang, Y. L., and R. L. Richards. 1994a. Coleoptile tiller development is associated with
fast early vigor in wheat. Euphytica 80: 119-124.
Liang, Y. L., and R. A. Richards. 1994b. Variation in seedling vigor among Chinese and
Australian wheats. In Proc. Int. Wheat Genetics Symp., 8th. Beijing,
20-25 July 1993.
Lindberg,, B. W. 1968. Statistical theory. Ed. 2. Macmillan, New York.
Lopez-Castaneda, C., and R. A. Richards. 1994. Variation in temperate cereals in
rainfed environments. II. Phasic development and agronomie characteriscs. Field
Crops Res. 37:5 l-62.
Lopez-Casteneda, C., R. A. Richards, and G. D. Farquhar. 1995. Variation in early vigor
between wheat and barley. Crop Sci. 35:472-479.

---

14.5
Ludlow, M. M., R. C, Muchow. 1990. A critical evaluation of traits for improving trop
yields in water-limited environments. Adv. in Agron. 43 : 107- 153.
Lynch, D. V., G. A. Thompson. 1984. Microsorna phospholipid molecular species
alterations during low temperature acclimation in Dunaliella. Plant Physiol.
74:193-197.
Ma, Y., F. A. Bliss. 1978. Tannin content and inheritance in common bean Croup Sci.
18:201-204.
Maiti, R. K., P. S. Raju, F. R. Bidinger. 1981. Evaluation of visual scoring for seedling
vigor in sorghum. Seed Sci. Technol. 9:613-622.
Mandoli, D. F., and W. R. Briggs. 1984. Fiber optics in plants. Sci. Am. 251:80-88.
Mann, J. A., E. E. Gbur, and F. R. Miller. 1985. A screening index for adaptation in
sorghum cultivars. Crop Sci. 25:593-598.
Maranville, J. W.,and M. D. Clegg. 1977. Influence of seed size and density on
germination, seedling emergence, and yield of grain sorghum.
Agron. J. 69;329-330.
Martin=, O., R. N. Curnow. 1992. Estimating the locations and the sizes of the effects
of quantitative trait loti using flanking markers. Theor. Appl. Genet. 85:480-488.
Mar-ther, K., J. L. Jinks 1971. Biometrical genetics. Corne11 Univ. Press. Ithaca, N.Y.
Illason,S. C., J. Lasschuit, and J. M.. Lasa. 1994. Inter&hn&ip of sorghum c&optile
morphology with emergence potential in crusted soils. Eu, J. Agron. 3( 1) 17-21.
McConnell, R. L., C. 0. Gardner. 1979. Selection for cold germination in two corn
populations. Crop Sci. 19:763-768.

---

1 4 6
McCovan, R. L., R. K. Jones, and D. C. 1. Pearke. 1985. Evaluation of no-till, tropical
legume, bey-farming strategy. Chapter 22. In R. C. Muchow (ed.) Agro-Research
for the semi-arid tropics. University of Queensland Press, St. Lucia, Australia.
McMillian, W. W., B. R. Wiseman, R. E. Burns, H. B. Harris, and G. L. Greene. 1972.
Bird resistance in diverse germplasm of sorghum. Agron. J. 64:821-822.
Melake B. A., S. H. Hulbert, L. G. Butler, and J. L. Bennetzen. 1993. Structures and
evolution of the genomes of Sorghum bicolor and Zea mays. Theor. Appl. Genet.
86:598-604.
Mendoza-Onofie, L., J. D. Eastern, C. Y. Sullivan, and W. M. Ross. 1979. Relations
between laboratory and field observations of sorghum seedlings grown under cold
temperature. In development of improved high-yielding sorghum cultivais. Ann.
Report. no.9 Univ. Nebraska, ARS, USDA, 137-145.
Mian, A. R., E. D. Nafziger. 1992. Seed size effects on emergence, head number, and
grain yield of winter wheat. J. Prod. Agric. 5:265-268.
Milic, B. L., S. Srdjan, and V. Nada. 1972. Leucern tannins; II, Isolation of tannin from
leucern, theirs nature and influence on the digestive enzymes in vitro. J. Sci. Food
Agri. 23 : 1157-
Mishra, A., B. S. Siradhana, A. Shivpuri. 1980. Phenols in relation to resistance of
sorghum to anthracnose. Philipp. Agit. 63 : 7 l-74.
Mock, J. J., S. A. Eberhart. 1972. Cold tolerance in adapted maize. Crop Sci. 12:230-
233.

---

147
Mock. J. J., and A. A. Bakri. 1976. Recurrent selection for cold tolerance in maize. Crop
Sci 16:230-233.
Mode, C. J. and H. F. Robinson. 1959. Pleotropism and the genetic variante and
covariance. Biometrics, 15: 5 18-537.
Monk, R. L. 1977. Characterization of grain sorghum emergence M. SC. Thesis, Texas
A and M University College Station, TX, 82 pp.
Monteith, J. L. 1987. Microclimatology in tropical agiculture, Volume 1, introduction,
methods and principles. Univ. of Nottingham, Overseas Development
Administration, Eland house, London, 202~~.
M’Ragwa, L. R. F., C. E. Watson, and L. M. Gourley. 1995. Selection response for
seedling root length and coleoptile length in Pearl millet. Crop Sci. 35: 1032-1036.
Nicholson, R. L., S. S. Kollipara, J. R. Vincent, P. C. Lyons, and G. Cadena-Gomez.
1987. Phytoalexin synthesis by sorghum mesocotyl in response to infection by
pathogenic and non-pathogenic fùngi. P. N. A. S. USA 84:5520-5524.
Nykaza, S. M. 1983. Coleoptile length evaluation in winter wheat. Ph.D. Diss.
Colorado State Univ. (Diss. Abstr. 44:3-2301B.
Nyquist.W. E. 1991. Estimation of heritability and prediction of selection response in
plant populations. Crit. Rev. Plant Sci. 10:235-322.
Ougham, H. J., J. M. Peacock, J. L. Stoddart and P. Sonam. 1988. High temperature
effects on seedling emergence and embryo protein synthesis in sorghum. Crop
Sci. 28:25 l-253.

---

148
Guassou, A., S. Baesinger, and A. Taylor. 1991. Genetic variation for Plant height and
coleoptile length in durum wheat. Agron. Abstr., AM, Madison, WL P. 109.
Pascual, M. J. V., G. J. Jellis. 1990. Factors influencing establishment in ficiafaba J.
Agri. Sci. Cambridge 115:57-62.
Pater-son, A. H., E. S. Lander, J. D. Hewitt, S. Peterson, S. E. Lincoln, S. D. Tanks@
1988. Resolution of quantitative traits into mendelian factors using a Com@te
linkage map of restriction fragment lenght polymorphisms. Nature 335:721-726.
Peacock, J. M. 1982. Response and tolerance of sorghum to temperature stress. In: L.
R. House et al. (ed.). Sorghum in the eighties. Proc. Int. Symp. Sorghum,
Patancheru, India. 2-7 Nov, 198 1. ICRISAT, Patancheru, India.
Pereira, M. G.,M. Lee,P Bramel-Cox, W. Woodmaan, J. Doebly, and R. Whitkus. 1994.
Construction of a RFLP map in sorghum and comparative mapping in maize.
Genome 37:236-243.
Perry, D. A. 1969. Seed vigor in peas (Pisum sativum L.). Roc. ht. S&d Test hoc.
341221-232.
c
Perry, D. A. 1980. Deterioration of barley seed and its effects on field performance .
P. 327-337. In P. D. Hebblethwaite (ed.) Seed production. Butterworths, London.
Price, M. L., L G. Butler. 1980. Tannins and nutrition, Agric. Exp. STN. Bull. N0272.
Dept of Biochem., Purdue LJniv., West Lafayette, In.
Radford, B. J., 1. M. Wood, C. H. Beavis, P. A. Walsh, A. M. Vieritz, W. J. L. Hazard, L.
J. Wade, P. J. Hughes, L. N. Robertson, J. R. Page, A. S. Wollin, And G. B.
Spakman. 1989. A survey of the establishment of commercial sorghum and

---

149
sunflower crops in the central highland of Queensland and analysis ofthe effècts
of level and evenness of establishment on grain sorghum. Qccasional publication
43. Australian Inst. of Agric Sci., Brisbane.
Radford, B. J.,and R. G. Henzell. 1990. Temperature affects the mesocotyl and
coleoptile length of grain sorghum genotypes. Aust. J. Agric. Res. 41:79-87.
Radford, .P. J. 1967. Growth analysis formulae. Their use and abuse. Crop Sci.
7:171-175.
Ragab, R. A., S. Dronavelli, M. A., Saghai-Maroof, and Y. G. Yu. 1994. Construction
of a sorghum RFLP linkage map using sorghum and maize probes. Genome
3’7:590-594.
Ram, C., O., Singh, R. P. S. Kharb, P. Kumari and T. P. Yadava. 199 1. Seedling vigor
in pigeonpea. Seed Sci. Technol. 19:627-63 1.
Rasmusson, J. M. 1935. Studies on the inheritance of quantitative characters in Pisum: 1.
Preliminary note on the genetics of flowering. Hereditas 20: 16 l_ 180.
Ray S. D., K N. Guruprasad, M. M. Laloraya. 1980. Antagonistic action of phenolic
compounds on abscissic acid-induced inhibition of hypocotyl growth. J. Exper.
Bot. 31(125):1651-1656.
Regan, R. L., K. H. M. Siddique, N. C. Turner, and B. R. Whan. 1992. Pote&1 for
im-easing early vigor and total biomass in spring wheat. 2. Characteristics
associated with early vigor. Aust. J. Agric. Res. 43:541-553.
Richards, R. A. 1987. Physiology and the breeding of winter-grown cereals for dry
areas. In ‘Drought tolerance in winter cereals’, (eds. J. P. Srivastava, E. Porceddu

---

150
E. Aceveddu, and S. varma) pp. 133-150. John Wiley ad sons LTD;
Chichester, U. K.
Rober-ts, E, H. 1972. Loss of viability and trop yields. p. 307-320. In E. H. Roberts
(ed.)
Vi,ability of seeds. Syracuse Univ. Press, Syracuse. NY.
Sax, K. 1923. The association of size differences with seed coat pattern and
pigmentation in Phaseolus vulgaris.
Genetics 8: 5 S2-560
Schaffer, M. A., R. L. fischer. 1988. Analysis of mRNAs that accumulate in response to
low temperature identifies a thiol protease gene in tomato. Plant Physiol. 87:43 I-
436.
Sharma, S., S. S. Sharma, V. K. Rai. 1986. Reversa1 by phenolic compounds of abcissic
açid-induced inhibition of in vitro activity of amylase from seeds of Triticum
aestivum L., New Phytol. 103:293-297.
Snedecor, G. W., and W. G. Cochran. 1989. Statistical Method. 8th ed, Iowa State
Univ. Press, Ames.
Soman, P., K.N. Rai, and F. R. Bidinger. 1989. Dwarfing gene effects on coleoptile
length in Pearl millet. Crop Sci. 29:956-958.
Spilde, L A. 1987. Navy bean seed vigor and field performance in relation to
microwave radiation. Agron. J. 79:827-830.
Storm, D. J. 1982. Effect of polyphenols on shoot and root growth and on seed
germination. Biologia plantarum 24 (1) : l-6.

---

151
Stuber, C. W. and M. D. Edwards. 1986. Genotypic selections for improvement of
quantitative traits in corn using molecular marker loti. Proc. 41th annual corn and
sorghum reseach conf. Am. Seed Trade ASSOC. 41 :JO-83.
Stuber, C. W, 1992. Biochemical and molecular markers in plant breeding. Annual Rev.
Plant Breeding. 9:37-61.
Suh, H. W., A. J. Casady, and R. L. Vanderlip. 1974. Influence of sorghum seed weight
on the performance of the resulting trop. Agron. J. 14:835-836.
Swaminathan, M. S., S. P. Kohli, and R. G. Anderson. 1956. Sonora 64, an early dwarf
wheat with high yield. Indian Fmg. 16:4-6.
Swanson, A. F., R. Hunter. 1936. Effect of germination and seed size on sorghum
stands. J. Amer. Soc. Agron. 28; 997-1004.
Tamir, M. and Alumot, E. 1969. Carot Tannins, growth depression, and levels of
insoluble nitrogen in the digestive tract of rats. J. Nutr. 100:573-
Tanksley, S. D. 1993. Mapping polygenes. Ann. Rev. Genet, 27:205-233.
Tanner, C. B.,and T. R. Sinclair. 1983. Efficient water use in trop production: research
or re-search? p. l-27. In H. M. Taylor et al. (ed.). Limitations to efficient water
use in trop production. ASA, CSSA, and SSSA, Madison, WI.
Tekrony, D. M.,T. Bustaman, D. B. Egli, and T. W. Pfeiffer. 1987. Effects of soybean
seed size, vigor and maturity on trop performance in row and hi11 plots. Crop Sci.
27:1040-1045.
Tekrony., D. M., B. D. Egli, and D. A. Wickham. 1989. Effect of cor-n seed vigor on no-
tillage field performance: II. Plant growth and grain yield. Crop Sci.

---

152
29:1040-1045.
Tekrony, D, M., D. B. Egli. 199 1. Relationship of seed vigor to trop yield: Areview.
Crop Sci. 31:816-822.
Thomas, G. L., and F. R. Miller. 1979. Base temperatures for germination oftemperate
and tropically adapted sorghum. p. 24, In Proc. Biennal Grain Sorghum Res. and
Utilization Conf, 1 lth. Feb. 2%Mar. 2 1979. Grain Sorghum Producers, ASSOC.,
Lubbock, TX.
Vanderlip, R. L. 1972. How a sorghum plant develops. Coop. Ext. Set-. C-447, Kansas
State Univ., Manhattan.
Vanderlip, R. L., F. E. Mockel, and H. Jan. 1973. Evaluation of vigor tests for sorghum
seed. Agron. J. 65;484-488.
Wade, L. J., G. L. Hammer, and M. A. Davey. 1993, Response of germination to
temperature among diverse sorghum hybrids. Field Crops Res. 3 1:295-308.
Wall, J.S., and W. M. Ross. 1970. Composition of sorghum plant ant grain. In Sorghum
Production and Utilization. J. S. Wall and C. W. Blessin, Ed. AV1 Publishing CO.
Wesport, Ct. p. 118-166.
Wanjari, K. B., and M. P. Bhoyar. 1980. Coleoptile Length in sorghum. Seed Sci.
Technol. 8: 169-174.
Watterson, J. J., and L. G. Butler. 1983. Occurrence of an unusual leucoanthocyafidin
and absence of proanthocyanidins in sorghum leaves. J. Agric. Food Chem.,
31:41-45.

---

153
Weerasuriya, Y. M. 1995. The contruction of a molecular linkage mal& maPPing of
quantitative trait loti, characterization of polyphenols, and screening of genotypes
for Striga resistance in sorghum. PH.D thesis Purdue University. pp. 178.
Whan, B. R., G. P. Carlton, and W. K. Anderson. 1991. Potential for increasing early
vigor and total biomass in spring wheat. 1. Identification of genetic improvements.
Aust. J. Agric. Res. 42:343-361.
Wheeler T. B. and R. H. Ellis. 1992. Seed quality and seedling emergence in onion
(AIlium Cepa L.). J. Hort. Sci. 67:319-332.
Whitkus, R., J. Doebly, and M. Lee. 1992. Comparative genome mapping of sorghum
and maize. Genet. 132:1119-l 130.
Wilson, G. L., P. S. Raju, and J. M. Peacock. 1982. Effect of soi1 temperature on
seedling emergence in sorghum. Indian J. Agric. Sci. 52: 848-85 1.
Willey, R.. W., and S. B. Heath. 1969. The quantitative relationships between plant
population and trop yield. Adv. Agron. 21:28 l-321.
Woodrt@ B. J., R. P. Cantrell, J. D. Axtell, L. G. Butler. 1982. Inheritance of tannin
quantity in sorghum. J. Heredity 73:214-218.
Wright, A. J., R. P. Mowers. 1994. Multiple regression for molecular-marker,
quantitative trait data from large F2 populations. Theor. Appl. Genet.
89:305-3 12.

---

154
VITA
Ndiaga Cisse was born on February 25, 1952, in Rufisque, Senegal. He had
completed his undergraduate studies at the ‘Agronomie Institute Nicolae Balcescu’
(IANB) in Bucharest, Romania, and obtained a M. S. degree in Agronomy at the
University of California at Davis in 1983. From 1983 to 1991, He worked as a cowpea
breeder for The ‘Institut Senegalais de Recherches Agricoles’ (ISRA). In August of 1991 ï
he joined the Agronomy Department at Purdue University, and completed his Ph.D.
requirements under the guidance of Dr. Ejeta, in December, 1995.

---