5.3.

. z .. $3.3.
L huh 3.... .. .. .7 E
.3 a. .

if
s... :i ...
c......:?§...

. . .63...
. (I .2: .
z Em... 3. ,

. . . l
. . 1...;
.. m... . , . _ . . .
§ , , fl... . mam
Afiflfiy , .
as“. 3...
nu.
«ma

545.431.

5: . vvfixuwnn .
is i
. . z. .
1. x: 2... .iu. ,
.....fi...s...z..

.~l..

m

.1.
.4.

5%.; :.
had-2.1...
.wk. 8. 11L:
. .

.Iafivdfiai :I .
ire.“-

:

i=5. .. dz...

.9!
3.0 A!

.
:53.
x hr .35.. .E
.
. «.L , .
5

. :2!

0:... 9
ii. 33,:
, :3. 4.2.3:!»
3:12.! 1511.:
. : :.|,..::.3I.|..s12
. S..!¢.)..5
)

g 3...
.. 3......
Vita}.
ii 3

 

1.3)...
.....v6- 20
.5... I .. .
. Err;
¢

,1 . _ Efifigugfi :3... 3.”.-.

 

THESIS
r. ,‘ .. ‘ a 'HDADV

WHVI “gal I State
University

 

 

 

This is to certify that the

thesis entitled
YIELD, SEED WEIGHT, AND CANNING QUALITY IN KIDNEY BEAN
(PHASEOLUS VULGARIS L.); AND RANDOM AMPLIFIED POLYMORPHIC
DNA (RAPD) MARKERS ASSOCIATED WITH CANNING QUALITY TRAITS

presented by

MARIA-CARMELA POSA MACALINCAG

has been accepted towards fulfillment
of the requirements for

M.S. degreein Plant Breeding and

 

 

Genetics

   

Major profe r

Date December 13, 2001

 

0.7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRCIDateOuepes-nts

 

 

ABSTRACT
YIELD, SEED WEIGHT, AND CANNING QUALITY IN KIDNEY BEAN
(Phaseolus vulgaris L.); AND RANDOM AMPLIFIED POLYMORPHIC DNA (RAPD)
MARKERS ASSOCIATED WITH CANNING QUALITY TRAITS
By

Maria-Carmela Posa Macalincag

Two recombinant inbred populations of kidney beans (Phaseolus vulgaris L.) -
‘Montcalm’ x ‘California Dark Red Kidney 82’ and ‘Moncalm’ x California Early Light
Red Kidney’ - were evaluated in six year-location combinations in Michigan, Minnesota
and North Dakota from 1996 to 1999. Heritability estimates were obtained for yield
(0.62 to 0.63), seed weight (0.58 to 0.69), and canning quality traits - appearance (0.83 to
0.85) and degree of splitting of processed beans (0.84 to 0.85). Positive correlations were
detected between yield and seed weight, and between APP and SPLT. Negative
correlations were detected between yield and APP, and yield and SPLT.

Two putative quantitative trait loci (QTL) for canning quality traits were
identified using eleven RAPD markers. The first region was tentatively mapped in
linkage group B8 of the bean genome. The alleles in this locus, which were associated
with desirable canning quality, appeared to be derived from Montcalm. The second
locus, associated with 4 markers, appeared to be derived from the non-Montcalm parents.

Population and environment-specificity were observed for the markers identified.

Copyrighted by
MARIA-CARMELA POSA MACALINCAG
2001

To my parents and my husband.

ACKNOWLEDGMENTS

I would like to express my heartfelt thanks to the following people for their part in
this work and in my graduate studies:
my advisor, Dr. George Hosfield, for his support, guidance and encouragement;
Dr. James Kelly, for the knowledge he imparted, and for sharing his enjoyment of plant
breeding;
Dr. Mark Uebersax, for his encouragement and guidance in the food quality aspects of
this work;
Dr. Kenneth Grafion, for providing the populations and some DNA samples, and for
conducting the experiments in North Dakota and Minnesota;
Shawna, Bernie and Robyn, for the many hours of assistance in the field and in the food
science lab;
Halirna, Kristin, Maeli and Judy, for their help in the marker work and for companionship
in the lab; Jerry and Norman, for their assistance in the field;
Kuya Gem, the Alociljas, and the Altamiranos, for the wonderful fellowship and
brotherly love;
the Macalincag and Viray families for welcoming me into their lives;
the MSU Filipino Club for all things “Pinoy”;
Tita Naty and Tito Noel, for their unselfish love and support;
Ma and Dad, Ate Grace and Mike, J onjon, Weng, Gabe and Justine, and Bern, for their
unconditional love and encouragement;
my husband, Bong, for his love and friendship, and for everything he is;
and finally, God Almighty, for without Him, there is nothing.

TABLE OF CONTENTS

LIST OF TABLES ............................................................................... viii
LIST OF FIGURES ............................................................................... xii
INTRODUCTION .................................................................................. 1

CHAPTER 1: YIELD AND SEED WEIGHT OF TWO KIDNEY BEAN
RECOMBINANT INBRED POPULATIONS .................................................. 6

Introduction ....................................................................................... 6
Review of Literature ............................................................................ 7
Yield and Yield Components .............................................................. 7
Genotype x Environment Interactions .................................................. 10
Materials and Methods ......................................................................... 11
Genetic Material 11
Field PlotProcedures 12
Statistical Analysis and Estimation of Heritability .................................... 14
' Results 16
High-yielding RILs in Populations l and 2 ............................................. 22
Heritability Estimates and Correlations Between Yield and Seed Weight 29
Discussion 32
CHAPTER 2: EVALUATION OF CANNING QUALITY 1N KIDNEY BEAN, AND
THE IDENTIFICATION OF RANDOM AMPLIFIED POLYMORPHIC DNA (RAPD)
MARKERS ASSOCIATED WITH CANNING QUALITY TRAITS 40
Introduction 40

Review ofLiterature 42

Canning Quality 43

vi

Use ofMarkersin Crop Improvement .
MaterialsandMethods .. .
EvaluationofCanning Quality...........................
Identification of RAPD Markers ..............................................
Statistical Analysis and Estimation of Heritability .........................
Results ...................................................................................
EvaluationofCanning Quality
Identification of Putative Markers for Canning Quality Traits
Discussion

50

6O

.61

............ 65

........... 7O

........... 74

74

93

119

134

LITERATURE CITED ............................................................................. 170

vii

LIST OF TABLES

Table 1. Parents of recombinant inbred populations 1 and 2 and varieties and breeding
lines used as checks in each population, and the years and locations in which they were
grown in the study .................................................................................. 13

Table 2. Significance levels for main effects and interactions for yield and seed weight of
Population 1 entries. Data analyses were according to individual experiments, years, and
environments (location and years confounded) ................................................ 17

Table 3. Significance levels for main effects and interactions for yield and seed weight of
Population 2 entries. Data analyses were according to individual experiments, years, and
environments (location and years confounded) ................................................ 18

Table 4. Data Analysis 1 - Yield and seed weight of Population 1 entries, including
parents, RILs and checks. Analyses were conducted individually for each
experiment ........................................................................................... 19

Table 5. Data Analysis 1 - Yield and seed weight of Population 2 entries, including
parents, RILs and checks. Analyses were conducted individually for each
experiment .......................................................................................... 19

Table 6. Data Analysis 2- Yield and seed weight of Population 1 entries, grown in
Michigan from 1996 to 1999, analyzed to compare individual years” 23

Table 7. Data Analysis 2 - Yield and seed weight of Population 2 entries, grown in
Michigan fi'om 1996 to 1999, analyzed to compare individual years ....................... 23

Table 8. Data Analysis 3 - Yield and seed weights of Population 1 parents and RILs
grown in Michigan, Minnesota and North Dakota from 1996 to 1999, analyzed to
compare year-location combinations, treated as environments .............................. 24

Table 9. Data Analysis 3 - Yield and seed weights of Population 2 parents and RILs
grown in Michigan, Minnesota and North Dakota fi'om 1996 to 1999, analyzed to

compare year-location combinations, treated as environments 24
Table 10. Yields of entries in Population 1 .................................................... 25
Table 11. Yields of entries in Population 2 .................................................... 26
Table 12. Seed weights of entries in Populations 1 and 2 ................................... 27

Table 13. Heritability estimates for yield and seed weight of the 75 and 73 RILs in
Populations l and 2, respectively, calculated fiom data combined over six
environments ....................................................................................... 29

viii

Table 14. Significant correlations between yield, seed weight and seed color in
Populations 1 and 2, planted in Michigan, Minnesota and North Dakota fiom 1996 to
1999 .................................................................................................. 30

Table 15. Scores for appearance of processed beans of the Population 1 RILs that were
used in the DNA bulks to screen RAPD primers for polymorphism ........................ 67

Table 16. Significance levels for main effects and interactions for canning quality traits
of Population 1 entries. Data analyses were according to individual experiments, years,
and environments (location and years confounded) ........................................... 75

Table 17. Significance levels for main effects and interactions for canning quality traits
of Population 2 entries. Data analyses were according to individual experiments, years,
and environments (location and years confounded) .......................................... 76

Table 18. Hydration coefficient and scores for appearance and degree of splitting of
processed beans of Population 1 entries. Data analyses were conducted individually for
each experiment (Analysis 1) .................................................................... 77

Table 19. Hydration coefficient and scores for appearance and degree of splitting of
processed beans of Population 2 entries. Data analyses were conducted individually for
each experiment (Analysis 1) .................................................................... 77

Table 20. Hydration coefficient and scores for appearance and degree of splitting of
processed beans of Population 1 entries. Data analyses were conducted to compare
' individual years in Michigan (Analysis 2) ..................................................... 81

Table 21. Hydration coefficient and scores for appearance and degree of splitting of
processed beans of Population 2 entries. Data analyses were conducted to compare
individual years in Michigan (Analysis 2) ..................................................... 81

Table 22. Hydration coefficient and scores for appearance and degree of splitting of
processed beans of Population 1 parents and RILs grown in Michigan, Minnesota and
North Dakota from 1996 to 1999. Analyses were conducted to compare year-location
combinations, treated as environments (Analysis 3) .......................................... 82

Table 23. Hydration coefficient and scores for appearance and degree of splitting of
processed beans of Population 2 parents and RILs grown in Michigan, Minnesota and
North Dakota from 1996 to 1999. Analyses were conducted to compare year-location
combinations, treated as environments (Analysis 3) .......................................... 82

Table 24. Scores for appearance and degree of splitting of processed beans of Population
1 parents and RILs that appeared in the ten highest scoring group of RILs in three and six
individual experiments with the mean scores of the experiment, all the RILs, the 10
highest scoring RILs, the parents, and the checks ............................................. 84

Table 25. Scores for appearance and degree of splitting of processed beans of Population
2 RILs that appeared in the ten highest scoring group of RILs in three, four and five
individual experiments with the mean scores of the experiment, all RILs, the 10 highest
scoring RILs, the parents, and the checks ................................. 89

Table 26. Heritability estimates for appearance and degree of splitting of processed
beans in Populations l and 2, using data from all environments combined ............... 92

Table 27. Significant correlations between canning quality traits, seed weight, and yield
in Population 1, planted in Michigan, Minnesota and North Dakota from 1996 to
1999 ................................................................................................. 94

Table 28. Significant correlations between canning quality traits, seed weight and yield
in Population 2, planted in Michigan, Minnesota and North Dakota from 1996 to
1999 ................................................................................................. 94

Table 29. Coefficients of determination (R2) of RAPD markers associated with scores
for appearance and degree of splitting of 75 RILs of Population I, planted in Michigan,
Minnesota and North Dakota, from 1996 to 1999 ........................................... 102

Table 30. Average scores for appearance of processed beans of two groups of Population
1 RILs that had the opposite alleles of the markers significantly associated with these
traits, and the genotypes of the parents (Montcalm and California Dark Red Kidney 82)
and RIL 118-90 ................................................................................... 104

Table 31. Average scores for degree of splitting of processed beans of two groups of
Population 1 RILs that had the opposite alleles of the markers significantly associated
with these traits, and the genotypes of the parents and line 118-90 ....................... 105

Table 32. Coefficients of determination (R2) for composites of RAPD markers
significantly associated with scores for appearance and degree of splitting of processed
beans of 75 RILs of Population 1, planted in Michigan, Minnesota. and North Dakota
fi'om 1996 to 1999 ................................................................................. 107

Table 33. Mean scores for appearance and degree of splitting of processed beans of
Population 1 RILs selected using groups of markers, analyzed per year-location
combination ....................................................................................... 109

Table 34. Average scores for appearance and degree of splitting of processed beans

of RILs of Population 1 that were selected using marker composites A and D, and
M1+Gl7 (linkage group M1 and OGl7.1300), and M1+AN 16 (linkage group M1 and
OAN16.3000) ...................................................................................... 110

Table 35. Coefficients of determination (R2) for RAPD markers in linkage group M2
that were significantly associated with scores for appearance and degree of splitting

of 73 RILs of Population 2, planted in Michigan, Minnesota, and North Dakota from
1996 to 1999 ....................................................................................... 112

Table 36. Average scores for appearance and degree of splitting of processed beans of
two groups of Population 2 RILs that had the opposite alleles of the markers significantly
associated with these traits, and the genotypes of the parents (Montcalm and California
Early Light Red Kidney) ......................................................................... 113

Table 37. Coefficients of determination (R2) of composites for RAPD markers
significantly associated with scores for appearance and degree of splitting of processed
beans of 73 RILs of Population 2, planted in Michigan, Minnesota, and North Dakota
from 1996 to 1999 ................................................................................. 115

Table 38. Mean scores for appearance and degree of splitting of processed beans of
Population 2 RILs selected using groups of markers, analyzed per year-location
combination ....................................................................................... 116

Table 39. Average scores for appearance and degree of splitting of processed beans
selected using marker composites A and D, and M1+Gl7 (linkage group M1 and
OGl7.1300), and M1+AN 16 (linkage group M1 and CAN 16.3000) ..................... 118

LIST OF FIGURES

Figure 1. Data Analysis 1 - box plots for a) yield (kg.ha") and b) seed weight (g.100‘I
seed) of Population 1 RILs, parents and checks, planted in Michigan, Minnesota, and
North Dakota from 1996 to 1999 ................................................................. 20

Figure 2. Data Analysis 1 - box plots of a) yield (kg.ha") and b) seed size (g.100") of
Population 2 RILs, parents and checks, planted in Michigan, Minnesota, and North

Dakota from 1996 to 1999 ........................................................................ 21
Figure 3. Processed beans of the cultivar Montcalm ......................................... 62
Figure 4. Processed beans of the cultivar California Dark Red Kidney 82 . . . .. 63
Figure 5. Processed beans of the cultivar California Early Light Red Kidney . . . . . . 64

Figure 6. Data Analysis 1 - box plots of scores for a) appearance and b) degree of
splitting of processed beans of Population 1 RILs, parents and checks, planted in
Michigan, Minnesota, and North Dakota from 1996 to 1999 ................................ 79

Figure 7. Data Analysis 1 - box plots of scores for a) appearance and b) degree of
splitting of processed beans of Population 2 RILs, parents and checks, planted in
Michigan, Minnesota, and North Dakota from 1996 to 1999 ................................ 80

Figure 8. Processed beans of recombinant inbred line 118—90, fi'om a cross between
Montcalm and California Dark Red Kidney 82 ................................................. 86

Figure 9. Processed beans of Population 1 recombinant inbred line 118-51, from a cross
between Montcalm and California Dark Red Kidney 82 . ................................... 87

Figure 10. Amplification of primer OG17, showing marker OGl7.1300, using DNA
from parents and some RILs of Populations 1 and 2 ........................................... 96

Figure 11. Linkage groups detected inPopulation 1 (MCM x CDRK 82). Ml-l
(0Y7.850, 0Q14.950, OP15.1150, OAGlO.1650, OA17.4000, 018.1600, OU20.1150),
total map distance = 25.9 cM; M2-1 (OAH17.700, OGl7.1300, OAN 16.3000,
OH18.1000), total map distance = 26.1 cM . ................................................. 97

Figure 12. Linkage groups detected in Population 2 (MCM x CELRK). Ml-2 (OY7.850,
OQ14.950, OP15.1150, OAG10.1650, OA17.4000, 018.1600, OU20.1150), total map
distance = 6.5 cM; M2-2 (CAI-117.700, OGl7.1300, OAN 16.3000, OH18.1000), total
map distance = 27.4 cM .......................................................................... 99

Figure 13. Amplification products of primers 018 and OU20, showing markers 018.1600
and OU20.1150, respectively .................................................................... 100

xii

INTRODUCTION

Dry bean (Phaseolus vulgaris L.) is an important staple in countries where animal
protein is limited or expensive. In some countries in Central and South America, and
Central and East Afiica, large quantities of beans are consumed and provide from one-
quarter to more than one-half of the dietary protein, and up to one-quarter of the energy
requirements (Shellie-Dessert and Bliss, 1991). Even in the United States where beans
are consumed mostly to add variety to diets, their contribution to dietary requirements is
appreciable. The considerable diversity for seed characteristics and eating preferences of
dry bean lead to its classification into 13 major market classes in the US. (U .S.
Department of Agriculture, 1982). Dark and light red kidney beans, two important
market classes, account for a sizable consumption. Light red kidney beans are used in
chili and chili products; dark red kidney beans are used mainly in salads and constitute a
significant component of restaurant salad bars, particularly in northern US. states.

Increased and stabilized yield over a range of environmental conditions is a
major goal of breeding programs. Newer varieties with improved characteristics are
always evaluated, and may be accepted or rejected commercially, with regard to their
yield potentials. In developing and testing dry bean breeding lines and cultivars, plant
breeders pay attention to data on yield performance, heritability of yield and components
of yield, correlations between yield and other traits of interest, and genotype x
environment interactions. Such information aids in planning a program to improve yield
and other economically important traits, and serves as a benchmark for the evaluation of

materials planted at different locations and in different years. Data on genotype x

environment interactions serve as a guide in estimating the most efficient allocation of
locations, years, and replications necessary for testing and selecting genotypes with
improved yield and other characteristics. Such data would also be useful indicators of the
amount of genetic variability available for selection.

In addition to yield, bean breeders also include canning quality improvement as an
important program objective. Although uncooked seeds may be bought in stores and then
cooked on the stovetop or in the oven, a large amount of the dry bean crop produced in
the US. is consumed as a pre-processed (canned) product. Commercial canners process
beans in plain water, brine, sugar solutions, tomato sauce, molasses or mixed vegetables
added during processing (Adams and Bedford, 1973; Deshpande et al., 1984). Regardless
of how beans are purchased by the consumer, “dry pack” or in tin cans, beans are
generally soaked or blanched, and must be cooked to render them palatable, inactivate
heat labile anti-nutrients, and permit the digestion and assimilation of protein and starch
(Deshpande et al., 1984). The steps used in preparing beans for eating cause structural
changes in cells that influence acceptance criteria by consumers and processors. The
criteria used by consumers include appearance, ease of preparation, wholesomeness,
mouth feel and texture. On the other hand, processors, although constrained by consumer
expectations, seek properties of beans that lend themselves to ease of commercial
preparation, processing efficiency, and a high can yield per unit weight of raw product
(W assirni et al., 1990). To this end, processors desire beans that exhibit rapid and
uniform seed expansion during soaking and/or blanching (Hosfield, 1998), and beans that

maintain intact seed coats coupled with a high water-holding capacity during processing.

The multiplicity of characteristics used to determine whether or not processed beans are
preferred and acceptable to processors and consumers is referred to as canning quality.

The evaluation of genetic materials for improved canning quality, in addition to
yield and other agronomic features, is necessary because a bean cultivar with poor
canning quality may be rejected by consumers regardless of how agronomically superior
it is (Kelly et al., 1998). On the other hand, selections with good canning quality are
discarded if they do not meet yield expectations. Incorporating the dimension of canning
quality improvement into a bean breeding program places a heavy burden on the breeder
to develop efficient selection practices.

Dry bean canning quality is more or less conceptual because its definition depends
on a multiplicity of variables of which no single one adequately describes the properties
preferred and required (Hosfield and Uebersax, 1990). Furthermore, canning quality
traits are controlled by quantitative trait loci (QTLs), resulting in continuous variation
among phenotypes (Hosfield et al., 1984b). The number of genes influencing canning
quality, and the influence of the environment on gene expression complicate the
identification of the effects of individual genes controlling canning quality traits, and
thus, makes it difficult to manipulate genes for improving genotypes.

Indirect selection using linked markers - marker-assisted selection (MAS) - is a
method that might increase selection efficiency within breeding programs. If a trait is
difficult and expensive to evaluate, under polygenic control, or highly influenced by the
environment (such as is the case for bean canning quality traits) MAS may be more
efficient than traditional selection methods based on phenotype (Dudley, 1993). The use

of markers to facilitate selection could shorten the breeding cycle in plants because the

breeder might be able to select a desirable trait in the early generations following
hybridization. Early-generation selection increases the efficiency of breeding programs
because unwanted genotypes can be discarded before they enter replicated field trials.
The use of MAS can also reduce costs, especially when conventional selection methods
require evaluating numerous genotypes or large samples. Various morphological and
molecular markers have been used in MAS for different crops. Before such markers can
be used, associations or linkages between these markers and the QTL of interest must be
identified (Dudley, 1993; Mildas et al., 1996).

Walters et al. (1997) identified random amplified polymorphic DNA (RAPD) as
molecular markers for canning quality in three populations of navy bean. Several RAPD
markers were found to be associated with the traits: visual appeal, texture and washed
drained weight of canned beans. In the published literature no studies on beans other than
that of Walters et a1. (1997-) have been reported where MAS has been used to select for
canning quality or molecular markers have been developed for this trait. Given the
genetic diversity between bean market classes, kidney beans may or may not possess the
same markers associated with the same traits that were identified for navy bean.

The importance of yield and processing quality in kidney bean, the paucity of
published information on both, and an interest in identifying RAPD markers associated
with canning quality traits prompted the present work. Information on the inheritance of
these traits and on the effect of the environment and genotype x environment interactions
were also sought, in order to provide insight into the amount of testing required to
characterize breeding lines reliably. This research is composed of two studies, the first of

which dealt with yield and seed weight of two recombinant inbred populations of kidney

beans planted in Michigan in 1996, 1997, 1998, and 1999, in Minnesota in 1996, and in
North Dakota in 1999. The specific objectives of Study 1 were to a) evaluate yield and
seed weight of two recombinant inbred populations of kidney beans planted in six
environments; and b) estimate heritabilities and pair-wise correlations between traits.

The objectives of the second study were to a) evaluate the general appearance and degree
of splitting of canned beans of two recombinant inbred populations of kidney bean
planted in six environments; b) estimate heritabilities and pair-wise correlations; c)
identify putative RAPD markers for canning quality; and d) determine whether markers
associated with canning quality are the same across market classes, specifically for kidney

beans and navy beans.

CHAPTER 1: YIELD AND SEED WEIGHT OF TWO KIDNEY BEAN
RECOMBINANT INBRED POPULATIONS

INTRODUCTION

Total dry bean production in the United States in 1999 was estimated at 33.3
million hundredweight (cwt). Light red kidney and dark red kidney beans respectively
accounted for about 1.4 million cwt and 1.0 million cwt of this production (U SDA-
NASS, 2000). Three of the principal bean-producing states are Minnesota, Michigan and
North Dakota. In 1999, Minnesota alone produced about 178,000 cwt and 597,000 cwt
of light red and dark red kidney beans, respectively (USDA—NASS and Minnesota
Department of Agriculture, 2000). Michigan’s total dry bean production in 1999 was
7.34 million cwt. Of these, 306,0000 cwt and 153,000 cwt were the light red kidney and
dark red kidney bean market classes, respectively (U SDA-NASS and Michigan
Department of Agriculture, 2000). North Dakota’s total dry bean production in 1999 was
8 million cwt (USDA-NASS and North Dakota Department of Agriculture, 2000).

Sustained efforts in yield breeding in dry bean require a continuous evaluation of
yield and its components. Breeders should also have some knowledge of the heritability
for yield, and the magnitude of genotype x environmental interactions influencing yield
in the populations in which they are selecting. The data obtained from the present study
will increase the published information available for dry beans in general and kidney
beans in particular. The present study on two kidney bean recombinant inbred
populations was conducted in Michigan, Minnesota and North Dakota. Kidney bean

production in these states contributes substantially to the bean canning industry. The

specific objectives of this study were to a) evaluate yield and seed weight of two
recombinant inbred populations of kidney beans planted in Michigan from 1996 to 1999,
in Minnesota in 1996, and in North Dakota in 1999; and b) estimate heritabilities and

pair-wise correlations between traits.

REVIEW OF LITERATURE

Yield in dry bean can be viewed in terms of three components: number of pods
per plant, average number of seeds per pod, and average seed size (Adams, 1967; Coyne,

1968; Nienhuis and Singh, 1985; Ranalli et a1, 1991).

Yield and Yield Components

The contribution of number of pods per plant to total seed yield has been reported
to be more important than the other two components (Coyne, 1968). Although yield
components are genetically independent, negative correlations among components exist
not only for beans but also for other crops (Adams, 1967). Negative correlations are
caused by developmental rather than genetic factors, and result in yield component
compensations and yield stability under various environmental stresses (Adams, 1967;
Al-Mukhtar and Coyne, 1981). Various authors have studied the correlations between
the yield components in beans planted in different environments, sometimes with varying
results (Coyne, 1968; Nienhuis and Singh, 1985; Nienhuis and Singh, 1988; Zimmerman
et al., 1984b).

Coyne (1968) found that most correlation coefficients among yield components

were low and positive in sign, indicating the possibility of increasing one without

reducing the other two. Nienhuis and Singh (1985) observed that seed weight had
negative phenotypic correlations with number of pods and seeds per pod, although no
association was found between the latter two. In the Nienhuis and Singh (1985) study,
both number of pods and seeds per pod were positively correlated with yield but seed
weight and yield were negatively correlated. These authors suggested that selection for
increased number of pods or seeds per pod should result in increased yield, but seed
weight would be reduced. In a later selection experiment, Nienhuis and Singh (1988)
found that the number of pods had significant negative correlations with both seeds per
pod and seed weight. Selection for number of pods appeared to reduce not only seed
weight but also yield and seeds per pod. Seeds per pod and seed weight were also
negatively correlated. The authors (Nienhuis and Singh, 1988) suggested that selection
for seeds per pod would increase yield only slightly, and reduce number of pods and seed
weight. Selection for seed weight would reduce seeds per pod, and increase number of
pods and yield only slightly. The conclusion drawn from this work was that selection for
seed yield per se appears to be the best approach for yield improvement in dry beans
(Nienhuis and Singh, 1988). In another study in beans, Ranalli et a1. (1991) reported
inverse relationships between the three yield components such that selection for one was
detrimental to the others. Seed yield was increased by simultaneous selection for the
yield components, using adequate selection intensity and a selection index composed of
more than one trait (Ranalli et al., 1991).

In dry bean, the three yield components, along with yield per se, have been
reported to be under the control of different modes of gene action. Pod number has been

reported as completely dominant (Coyne, 1968), partially or almost completely dominant

(Sarafi, 1978) and with additive effects (Nienhuis and Singh, 1988). Sarafi (1978)
reported seeds per pod as partially or nearly completely dominant. Nienhuis and Singh
(1988) and Singh et al. (1991) found additive variance more significant than non-additive
variance for the trait. Mean seed weight was observed to be influenced by additive
effects (Coyne, 1968; Nienhuis and Singh, 1988; Singh et al., 1991) and partially or
nearly completely dominant (Sarafi, 1978). Nienhuis and Singh (1988) and Singh et a1.
(1991) found additive genes to be significant for yield per se. Zimmerman et al. (1985)
reported additive and dominance gene action, along with epistasis, as significant for yield
in some crosses.

Estimates of heritabilities reported for yield and yield components in beans
ranged from very low to high. In the cross Great Northern 1140 x PI 165078, low
heritability estimates were obtained for total seed yield (0.09 to 0.11) and for each of the
three yield components (—0.01 to —0.08) (Coyne, 1968). Sarafi (1978) found narrow
sense heritability estimates to be 29% for pods per plant, 38-42% for seeds per pod and
33-37% for 100-seed weight in a cross between Iranian and American bean cultivars
evaluated in the F2 and F3 generations. Zimmerman et al. (1984b) reported broad sense
heritabilities for yield to range from 0.21 to 0.23, number of pods to range from 0.63 to
0.86, seeds per pod to range from 0.81 to 0.90 and 100-seed weight to range from 0.97 to
0.99 in beans. For beans of Middle-American origin, Nienhuis and Singh (1988)
estimated narrow sense heritabilities to be 0.21 :t 0.13 for yield, 0.20 i 0.13 for number
of pods, 0.57 d: 0.13 for seeds per pod, and 0.74 :t 0.15 for seed weight. For a group of
genotypes mostly of Andean origin, Singh et a1. (1991) estimated narrow sense

heritability values to be 0.43 d: 0.19 for yield, 0.49 :L- 0.20 for number of pods, 0.63 :1: 0.21

 

for number of seeds, and 0.76 :h 0.23 for 100-seed weight. Other authors reported the
following broad-sense heritability estimates for seed yield: 0.90 (Scully et al., 1991), 0.42
:1: 0.07 to 0.49 :t 0.04 (Singh and Urrea, 1995) and 0.19 d: 0.17 to 0.50 21:0.16 (Welsh et al.,

1995).

Genotype x Environment Interactions

The presence of genotype x environment interactions is the reason that the
performance of any genotype relative to another grown in the same environment is
inconsistent. These interactions result in either a change in the ordering of the genotypes
(change in rank) fi'om one environment to another or to changes in the degree of
difference between them without changing their relative order (change in variance) (Hill,
1975). Genotype x environmental interactions are especially important if the relative
order of the genotypes changes (F ehr, 1987).

In tropically adapted gerrnplasm, Beaver et a1. (1985) observed that the magnitude
of the genotypic variance was similar to the variances of genotype x environment
interactions, indicating that these interactions are important factors to consider and that
testing must be done at several locations to obtain a precise estimate of yield. Likewise,
Nienhuis and Singh (1988) reported significant interactions in their work with 80

genotypes, which were mostly small-seeded and of Middle-American origin.

10

MATERIALS AND METHODS

Genetic Material

Two recombinant inbred populations of kidney bean provided the experimental
materials on which yield and seed weight were evaluated in the present study. These
populations were derived from ‘Montcalm’ (MCM), ‘California Dark Red Kidney 82’
(CDRK 82) and ‘California Early Light Red Kidney’ (CELRK). MCM is a dark red
kidney bean with a Type I grth habit, and was released in 1974 by the Michigan
Agricultural Experiment Station (Copeland and Erdmann, 1977). MCM is tolerant to
halo blight disease caused by Pseudomonas syringae pv. phaseolicola (Burkholder)
Young et al., matures in 90-100 days from planting, and has excellent canning quality.
CDRK 82 is a Type I growth habit dark red kidney bean released in 1989 by the
California Agricultural Experiment Station (CABS). CDRK 82 is resistant to bean
common mosaic virus (BCMV) and has good yield potential. CELRK was released in
1989 by the CAES. CELRK has a Type I growth habit, resistance to BCMV, and good
yield potential. CDRK 82 and CELRK mature in about 90 days and 80 days,
respectively, near Chico and Linden, California (Peterson, California Crop Irnprov.
Assoc., personal communication, Nov. 6, 2000).

Population 1, derived flour a cross between MCM and CDRK 82, comprised of
75 dark red kidney bean recombinant inbred lines (RILs). Population 2 comprised 73
RILs and was derived flour a cross between MCM and CELRK. The crosses were made
in 1991 by K.F. Grafton of the North Dakota Experiment Station. The protocol used to
develop the RH.s of each population was as follows: The initial selection of RILs was

made in the F2 generation. F2 plants were advanced in the greenhouse until the F6

11

generation, using the single-seed descent (SSD) procedure. Seed from F6 plants were

bulked, and the seed increased in the field until the F3 generation.

Field Plot Procedures

The 75 and 73 RILs of Populations 1 and 2, respectively, the two parents of each
population, and check genotypes (Table 1) were planted to conform to a 9 x 9 balanced
lattice (Cochran and Cox, 1968) for each population. The F63, F69, F630 and F6," RILs
of each population were planted in 1996, 1997, 1998, and 1999 on a McBride Sandy
Loam (coarse-loamy, mixed, frigid Alfic Fragiothods) at the Montcalm Research Farm
near Entrican, MI. Each population was planted in a separate experiment and replicated
two times, except in 1996, when the experiments were planted in three replications. The
entries were planted in two-row plots, 6.1 m long and spaced 0.5 m apart. Within-row
spacing was 7.6 cm. Herbicide and fertilizer applications were made following
recommendations for commercial bean production for each respective year. The
harvested area was 4.6 m2. The plants were harvested by hand from 1996 to 1998 and
threshed using a stationary plot thresher. In 1999, the plots were harvested mechanically
and threshed using a Hege 140 Plot Harvester (Hege Equipment, Inc.).

Populations 1 and 2 were grown in. Hubbard soil (sandy, mixed, fiigid, Entic
Hapludolls) in Perham, MI in 1996 (F 6,3) and in Gardena soil type (coarse-silty, mixed,
superactive, frigid, Pachic Hapludolls) in Erie, ND in 1999 (Fm 1). Table 1 gives the
details of the composition of entries for each year and location. These entries were
planted in two-row plots, 6.1 m long and spaced 0.8 m apart. The harvested area was 6.0
m2. In this location, the plants were harvested by hand and threshed using an Almaco

stationary plot thresher. After harvest at both the Minnesota and North Dakota sites, the

12

Table 1. Parents of recombinant inbred populations 1 and 2 and varieties and breeding
lines used as checks in each population, and the years and locations in which they were

gown in the study.

Variety or breeding line

Population 1"
MCMdc

CDRK 82‘"

Isles

Red Hawk

K9320] (Montcalm/37-l6)

K94202 (Sacramento/189021)
K97305 (Red Hawk/Drake)

K97309 (Red Hawk/K93644)
K90122 (Iassen/Isabella/Montcalm)

Population 2c

MCMdc

CELRK“

CDRK 82"

Isles

Chinook

Redhawk

K93621 (CELRK/Chinook)“
K93629 (CELRK/Chinook)8
K93653 (Chinook/CELRK)“
K93654 (Chinook/CELRK)“
K94515 (K89829/K88401)
Chinook2000

K97503 (Red Hawk/CELRK)“
K97504 (Red Hawk/Foxfire)

Year and locationa

Mich Minn Mich Mich Mich NDak

1996 1996

*1! 1} 'I'

1997 1998 1999 1999

1*
‘I'
'I
1}

*****

 

' *- indicates that variety or breeding line was grown in that particular year and location
b Population 1: Montcalm x California Dark Red Kidney 82
° Population 2: Montcalm x California Early Light Red Kidney

d Parents of the population
‘ MCM - Montcalm

‘ CDRK 82 - California Dark Red Kidney 82
‘ CELRK - California Early Light Red Kidney

13

seeds were hand-cleaned to remove split, damaged and diseased seeds. The seeds were
stored at room temperature (~22 °C) until sample preparation and analysis. The yield
(kg-ha") and lOO-seed weight (g) of each entry were recorded at constant moisture of

18%.

Statistical Analysis and Estimation of Heritability

All data were subjected to an analysis of variance (ANOVA) appropriate to a
randomized complete block design, with genotypes as random effects, and years and
environments (year-location combinations) as fixed effects. The SAS program proc glm
(SAS Institute, Cary, NC, 1998) was used to analyze data. Significance levels were set
at or = 0.05. Since the data from the study were not balanced in the sense that
experiments were grown in Michigan, Minnesota and North Dakota in each of the years
1996, 1997, 1998 and 1999, analyses were conducted according to the following groups:

Analysis 1 - separate analysis for each experiment i.e., MI-1996, MI-1997, MI-1998,
and MI-1999; MN-1996; and ND-1999.

Analysis 2 - combined data for Michigan over the years, 1996, 1997, 1998, and 1999.

Analysis 3 - combined analysis of all experiments such that combinations of years
and locations were treated as environments; only the parents and RILs of each
population were included in this analysis.

Box-plots of the data in Analysis 1 were constructed to provide a visual
comparison of the ranges, means and median values in the different environments. Box-
plots are interpreted as follows (Schabenberger, 1997):

a) mean - represented by (+)

b) median value - located by the line dissecting the box

14

c) first (Q,) and third (Q3) sample quartiles - determine the dimensions of the box. In an
ordered data set, 25% of all observations are smaller and 75% are larger than Q1; 25% are
larger and 75% are smaller than Q3. The difference between Q, and Q3 is called the inter-
quartile range (IQR).
d) whiskers - represent values within 1.5 x IQR fiom each end of the box
e) extreme values or outliers - Mild outliers (o) are observations beyond the whiskers but
less than 3 x IQR fi'om the respective end of the box. Extreme outliers (*) are
observations more than (3 x IQR) fi'om each end of the box.

For the estimation of heritability, two replications of the data fi'om the RILs in
Analysis 3 were used. Heritability was estimated for yield and seed mass on a progeny

mean basis (Fehr, 1987) as follows:

2

 

H2 = __L_“2 = OJ
oz. (32¢er + Ozzy/V + 0'23
where: 022 = genotypic variance
2

o . = total variance among RILs compared in r replications and v
environments (r = 2, v = 6)
02., = experimental error
°st = variance due to genotype x environment interactions
Confidence intervals for heritability estimates were derived according to Knapp et
a1. (1985). Correlations among the traits for each environment were also determined
using the program proc can in SAS (SAS Institute, Cary, N.C, 1998). To determine
correlations of seed color with the yield and seed weight, numerical values were
assigned, as follows: 1 - light red seed color, 2 — non-commercial seed color (a mixture

of light and dark red), and 3 — dark red seed color.

Images in this thesis are presented in color.

15

RESULTS

Genotypic effects were significant for both yield and seed weight in all data
analyses for Population 1 (Table 2). Except for yield in Mich-1998 (Analysis 1),
genotypic effects for Population 2 were significant for yield and seed weight in all
experiments (Table 3). Analyses of the Michigan combined data (Analysis 2) for both
populations showed significant year effects for yield and seed weight, which led to
significant interactions between years and entries (Tables 2 and 3). In Analysis 3 in both
populations (years and locations treated as environments), the genotype, environment and
genotype x environment effects were significant for both traits (Tables 2 and 3).

For the six experiments in Populations 1 and 2 (Analysis 1), the highest yields
were obtained in Mich-1999: 3197 kg-ha" in Population 1 (Table 4) and 3467 kg-ha" in
Population 2 (Table 5). These data are displayed pictorially in the box plots in Figures 1a
and 2a. The yield of the lowest yielding entries in Population 1 in Mich-1999 was higher
than that of most of the entries in North Dakota in the same year (Figure 1a). However,
due to the high amounts of variability in Minn-1996, Mich-1997 and Mich-1998, several
outliers in these environments had yields comparable to some of the highest yielding
entries in Mich-1999 (Figure la). In Population 2, mean yield was highest in Mich-1999;
no extreme differences in variability were observed among the environments (Figure 2a).
The yields in Mich-1996 for Population 1, and in NDak-1999 for both populations were
generally low for kidney beans (Tables 4 and 5). Seed weight was highest in Mich-1999
and lowest in NDak-1996 year for both populations (Table 4, Figure 1b; Table 5, Figure
2b). Seed weight observed in Minnesota and in North Dakota was generally low for

kidney beans. Ranges for seed weight, though variable, were somewhat similar across the

16

Table 2. Significance levels for main effects and interactions for yield and seed weight
of Population 1 entries. Data analyses were according to individual experiments, years,
and environments (location and years confounded).
Source of Variation Yield Seed weight
(kg.ha")' (_g.100 seed")'

Data analfiis number and location-year description
1 - Individual experiments

Michigan 1996: Genotype M n
Minnesota 1996: Genotype ** at:
Michigan 1997: Genotype M n
Michigan 1998: Genotype ** on
Michigan 1999: Genotype H u
North Dakota 1999: Genotype ** n

2 - Michigan data combined (1996, 1997, 1998, 1999)

Genotype it tap
Year , it **
Genotype x Year in u
3 - Locations and Years Confounded, and Treated as Environments
Genotype ** n
Environment u u
Genotype * Environment “I u

 

' ** - Significant at 0.05 level of significance

1?

Table 3. Significance levels for main effects and interactions for yield and seed weight
of Population 2 entries. Data analyses were according to individual experiments, years,
and environments (location and years confounded).

Source of Variation Yield Seed weight
(kg.ha")“ (5.100 seed")'

 

 

Data analysis number and location-year description
1 - Individual experiments

Michigan 1996: Genotype I" u:
Minnesota 1996: Genotype I" on
Michigan 1997: Genotype ** in
Michigan 1998: Genotype us In
Michigan 1999: Genotype H in
North Dakota 1999: Genotype " u

2 - Michigan data combined (1996, 1997, 1998, 1999)

Genotype " **
Year * O t *
Genotype x Year " **

3 - Locations and Years Confounded, and Treated as Environments

Environment ** "”"
Genotype * Environment ‘* **

 

‘ “ - Significant at 0.05 level of significance

18

Table 4. Data Analysis 1 - Yield and seed weight of Population 1 entries, including
parents, RILs and checks. Analyses were conducted individually for each experiment.

 

 

Leld Seed weight
Environment Mean Coefficient of Mean Coefficient of
(kg.ha") variation (%) (g. 100 seed') variation (%)
Mich (1996) 2615 21.1 56.2 6.1
Minn (1996) 2107 23.5 53.7 7.7
Mich (1997) 2345 19.9 61.9 4.8
Mich (1998) 2602 13.7 58.2 4.6
Mich (1999) 3197 11.6 63.4 3.5
NDak (1999) 1590 18.4 44.9 7.4

 

Table 5. Data Analysis 1 - Yield and seed weight of Population 2 entries, including
parents, RILs and checks. Analyses were conducted individually for each experiment.

 

 

m S e wei t
Environment Mean Coefficient of Mean Coefficient of
(kg.ha") variation (%) (g.100 seed') variation (%)
Mich (1996) 3359 16.8 61.9 4.4
Minn (1996) 2414 17.5 58.1 6.1
Mich (1997) 2199 15.5 64.4 5.8
Mich (1998) 271 1 16.6 59.5 4.7
Mich (1999) 3467 13.9 63.4 3.3
NDak (1999) 1491 21.4 47.6 6.7

 

l9

 

 

A. Yield (kgoha'l)

 

 

 

 

 

 

 

 

 

 

4500 e
Legend:
I 0 — extreme outliers
4000 6 I l-mildoutliers
I I I I am...
I I I I " ‘-median
3500 . I . I . ..... ,
I I I I I
I I I .I-Qoo.
I I I
3000 o I I * """ ’ I I
I o ooooo O I I 9 """" 9 I
I I I I I I I
§ ..... Q I I ...+-.. I I
2500 o I l I I I I I I
I I I 0 I I I I I
.3....' ' ..... ' 4 ----- O I I
I 9 l I I I I ’ """ ’
2000 o | I I I I I I I I
. ..... e I I o ----- e I I I
.IIOOO' I I I I 0 0 IIIII .
o I I I I I 0 l
1500 o 0 0 ----- 9 I I I I
I I I I
I I I I
‘ I I I I
1000 o I I I ’ """ ’
I I
I
I
500 o I I
I
I
o , ............ . ...................... . ........... . ...................... . ........... Environment

Mich96 Minn96 Mich97 Mich98 Mich99 NDak99

B. Seed size (g-100 seed")
so .

 

I
I
I o
70 9 0 I 0 I
I . ----- . l I
I I I I I e ----- o
I I 0...... I 0......
60 O O ----- O I I I e ----- o o ooooo o
'00-... O ..... O 9 ----- O .009... I I
I O I 0 ..... 0 I o ..... Q I I
. ----- . l + I I I I
50 o I o ----- 0 I I I
I I I o ----- o
0 I I 00.0...
I o ----- o
‘0 o 0 I I
I
I
30 o
20 o
0
1o , .............................................................................. Environment

Mich96 Minn96 Mich97 Mich98 Minn99 NDak99

Figure 1. Data Analysis 1- Box plots for a) yield (kg ha ) and b) seed weight (g- 100
seed ) of Population 1 RILs, parents and checks, planted in each environment.

20

 

A. Yield (kg-ha'l) Legend:

 

 

 

 

 

 

0 - extreme outliers
6000 9 |- mild outliers
+ - mean
° " * - median
sooo . I ,
I I
I |
I I
9 ----- 9 I I 9 ----- 9
o ..... t I I I o..,...
I 0 I I I I I I
3(’°‘) O O ----- 9 I I Q ..... Q 9 ..... Q
I O ----- O I '-.§--. I I
I 0.09... O ----- O I I I I
I I I 0.09... 9 ..... o I g ..... 9
2000 9 I 9--]--9 | I I I I
I Q ..... Q .00....
I I I 0 I 9 I
I I I I I
I I 9 ----- 9
I
° I
o . I
....................... ............ . Environment

Mich96 Minn96 Mich97 Mich98 Mich99 Ndak99

B. Seed size (g-100 seed")
I
so .

 

0
o I |
I I 0 I
I I 0 |
7() 9 | | I I |
I I 0 °°°° 9 I 9 ooooo 9
. ..... . I .OIOOI' I I I
.009... 9 ooooo O I O ----- Q OCIOII.
60 9 9 ----- 9 | I 9 ----- 9 Our-9... 9 ----- 9 I
I '--*'-' I I I I I
I 9 ----- 9 I O ----- 0 I I
I I I I I I
5" s l l I 9 ----- o
I o I 0 00.0...
I 0 9 ----- 9
o o I
4() 9 I
I
I
0
3!) 9
0
2C) 9
................................................................... .........-.- Environment

Mich96 Minn96 Mich97 Mich98 Mich99 Ndak99

Figure 2. Data Analysis 1 - Box plots of a) yield (kg-ha") and b) seed size (g-100 seed")
of Population 2 RILs, parents and checks, planted in each environment.

21

six environments for both populations. These results were reflected in the analyses based
on the four years of planting in Michigan (Analysis 2) (Tables 6 and 7), and on the
analyses which included only the parents and RILs (Analysis 3) (Tables 8 and 9).
High-yielding RILs in Populations l and 2

When the two parents of Population 1 were considered, MCM had yields higher
than that of CDRK 82 in three environments and CDRK 82 had higher yields in the other
three (Table 10). In Population 2, each of the two parents, MCM and CELRK, also had
higher yields than the other in three environments (Table l 1). In both populations, the
mean yield and seed weight of all the RILs did not exceed the mean yield and seed weight
of their parents (Tables 10, 11 and 12). However, in each environment, the RILs with the
ten highest yields exceeded the parent with the higher yield. Differences were significant
in some environments.

The 10 highest yielding RILs in Population 1 had a higher mean yield than the
check entries. For example, in Mich-1999 (Table 10), the 10 highest yielding RILs had a
mean yield of 3718 kg-ha", compared to the mean yields of MCM and CDRK 82 (3294
kg-ha"), all the RILs (3187 ltgha"), and the check varieties (3339 ltg-ha"). Differences
were significant only for NDak-1999. The mean seed weight of the 10 RILs with the
highest yields in each environment, on the other hand, was not consistently higher than
the seed weights of the parents (Table 12).

In Population 1, several RILs in the group with the ten highest yields had yields
higher than or comparable to the yield of either parent in more than two environments

(Table 10). One RIL of Population 1, 118-82, was common to the group with the 10

22

Table 6. Data Analysis 2 - Yield and seed weight of Population 1 entries, gown in
Michigan fiom 1996 to 1999, analyzed to compare individual years.

 

 

 

Year Yield“l Seed weightll
(kgha'l) (3.100 seedb
1996 2622 b 56.4 d
1997 2339 c 62.0 b
1998 2590 b 58.1 c
1999 3201 a 63 .4 a
Mean combined over years 2680 59.6
Coefficient of variation (%) 17.5 5.0

 

' - Means with the same letter are not sigrificantly different by Fisher's LSD (0.05).

Table 7. Data Analysis 2 - Yield and seed weight of Population 2 entries, gown in
Michigan fi'om 1996 to 1999, analyzed to compare individual years.

 

 

 

Year Yieldi Seed weight'I
Giana") (3.100 seed")
1996 3337 b 62.1 c
1997 2197 d 64.6 a
1998 2722 c 59.9 d
1999 3474 a 63.5 b
Mean combined over years 2977 62.5
Coefficient of variation (%) 16.4 4.5

 

‘ - Means with the same letter are not siglificantly different by Fisher's LSD (0.05)

23

Table 8. Data Analysis 3 - Yield and seed weights of Population 1 parents and RILs
gown in Michigan, Minnesota and North Dakota fi'om 1996 to 1999, analyzed to
compare year-location combinations, treated as environments.

 

 

 

Environment Yieldfi Seed weight'l
(kg.ha") (g.100 seed")
Mich 1996 2619 b 56.4 d
Minn 1996 2107 d 53.7 e
Mich 1997 2336 c 62.0 b
Mich 1998 2581 b 58.0 c
Mich 1999 3190 a 63.4 a
NDak 1999 1608 c 45.0 f
Mean combined over
environments 2424 56.4
Coefficient of variation (%) 18.8 5.7

 

 

' - Means with the same letter are not siglificantly different by Fisher‘s LSD (0.05).

Table 9. Data Analysis 3 - Yield and seed weights of Population 2 parents and RILs
gown in Michigan, Minnesota and North Dakota from 1996 to 1999, analyzed to
compare year-location combinations, treated as environments.

 

 

Environment Yield1 Seed weight'

(kg.ha") (5.100 seed")
Mich 1996 3321 b 62.1 c
Minn 1996 2413 d 58.1 e
Mich 1997 2199 c 64.5 a
Mich 1998 2720 c 60.0 d
Mich 1999 3458 a 63.6 b
NDak 1999 1500 f 47.7 f

 

Mean combined over
environments 2424 56.4
Coefficient of variation (%) 18.8 5.7

 

 

‘ - Means with the same letter are not sigrificantly different by Fisher’s LSD (0.05).

24

Table 10. Yields of entries in Population 1.
Environment
Entry Mich Minn Mich Mich Mich NDak
1996 1996 1997 1998 1999 1999

RIL that was common to the goup with the 10 highest fields in five of the six

environments”
118-82 3533 3181 3119 3268 3811 -"
Yield (kgs.ha")

 

 

RIL that was copppon to the gmup with the 10 highest fields in four of the s_ix_

environments”
118-46 3391 2986 3240 -" -° 2383
Yield (kgs.ha'1)

 

 

RILs that were common to the gmup with the 10 highest fields in three of the six

environments”
118-33 3’ 2801 3342 3014 -" -°
118-84 3670 -" -° -" 3604 2508
Yield (kgs.ha'l)

 

 

Parents of Population 1
CDRK 82' 2613 1273 2715 2670 2875 99

Montcalm 2790 2499 1381 2649 3713 2159
Yield (kgs.ha")

 

 

Means of the experiment, checks, parents, all 75 RILs in Population 1, and 10 highest
fielding RILs

Experiment 2615 2107 2345 2602 3197 1590
Check varieties 2598 -" 2526 2993 3339 1244
Parents 2702 1886 2048 2660 3294 1129
All RILs 2617 2105 2344 2579 3187 1620
Ten highest yielding RILs 3224 3034 3213 3078 3718 2436
LSD (0.05) 891 1000 928 709 738 581

cv (%) 21.1 23.5 19.9 13.7 11.6 18.4

 

 

Yield (kgs.ha")

 

' CDRK 82 - California Dark Red Kidney 82
" - only the yields where the RILs were among the ten highest-yielding lines are shown

25

Table 11. Yields of entries in Population 2.

 

Accession Seed colora

Mich

1 996

Minn

1 996

W
Mich Mich
1997 1998

Mich

l 999

NDak

1 999

 

RILs that were common to the goup with the 10 highest fields in four of the six

environments”
119-21 Light red
1 19-32 Dark red

C

C

3191
3069

 

C

C

Yield (kgs.ha")

3176
3248

4086
3991

2274
2293

 

RILs that were common to the gmup with the 10 highest fields in three of the six

environmentsb
1 19-17 Light red
119-50 Light red
1 19-60 Non-commercial
119-70 Light red
119-79 Light red

Parents of Population 2
CELRK” Light red
Montcalm Dark red

C

C

 

3801 3006
- 3076
-c 3245

4621 -c

3823 1964

3578 2120

 

2842 3128
-0 _c
-° 3088
-° 3312
-0 _c
Yneh1(kgsrurfil
2610 3242
1580 2610

Yield (kgs.ba")

C

 

4153 -
-c 2294
-° 2532

4342 -°

4265 2337

3214 436

3563 2546

 

Mm pf the pxpph'ment, check_s, pareng, all 75 RM in Popplation l, and 10 highest
fielding RILs, and values for LSD and CV

Experiment mean

Check varieties

Parents

All RILs

Ten highest yielding RILs
LSD (0.05)

CV (%)

3359
3836
3700
3311
4014
913
16.8

2414
2459
1806
2423
3130
853
17.5

 

2199
2204
2095
2201
2849
679
15.5

Yield (kgs.ha")

271 1
2597
2926
2714
3177
896
16.6

3467
3582
3388
3459
4145
961
13.9

1491
1379
1491
1500
2453
634
21.4

 

 

' Non-commercial seed color: a mixture of dark and light red

" CELRK - California Early Light Red Kidney

° - only yields in the environments where the RILs were among the 10 highest-yielding

lines are shown

26

Table 12. Seed weights of entries in Populations l and 2.

 

 

Environment
Accession Mich Minn Mich Mich Mich NDak
1996 1996 1997 1998 1999 1999
Population 1
Parents of Population 1
MCM 55.1 57.2 58.4 57.9 62.8 47.5
CDRK 82 57.8 49.1 54.2 55.6 65.7 42.1

 

 

Seed weight (g.100 seed")

Means of the experiment, checks, parents, all 75 RILs in Population 1, and 10 highest

 

 

fielding RILs
Experiment 56.2 53.7 61.9 58.2 63.4 44.9
Check varieties 53.2 - 59.9 61.7 63.2 42.9
Parents 56.4 53.1 56.3 56.7 64.2 44.8
All RILs 56.4 53.7 62.2 58.1 63.4 45.0
Ten Highest yielding RILs 58.8 57.7 66.7 58.3 63.2 48.4
LSD (0.05) 5.5 8.4 5.9 5.4 4.4 6.6
cv' (%) 6.1 7.7 4.8 4.6 3.5 7.4
Seed weight (g.100 seed")

Population 2

Parents of Population 2

MCM 62.8 56.0 69.0 56.3 61.9 46.9
CELRK” 64.0 50.2 61.6 64.5 60.9 44.2

 

 

Seed weight (g.100 seed")

Mm of the experimenh checks, parents, all 73 RILs in Population 2, and 10 highest

fielding RILs, and values for LSD and CV

Experiment mean 61.9 58.1 64.4 59.5 63.4
Check varieties 58.9 60.0 62.6 53.0 61.8
Parents 63.4 53.1 65.3 60.4 61.4
All RILs 62.1 58.1 64.5 60.0 63.6
Ten Highest yielding RILs 65.9 60.7 65.3 58.0 64.9
LSD (0.05) 4.4 7.2 7.5 5.5 4.1

CV' (%) 4.4 6.1 5.8 4.7 3.3

 

47.6
46.0
45.6
47.7
52.8
6.4
6.7

 

Seed weight (g.100 seed")

 

'CV - Coefficient of variation
° CELRK - California Early Light Red Kidney

27

highest yields in five of the six environments. RIL 118-46 was common to the goup
with the 10 highest yields in four of the Six environments. Moreover, in these
environments, RILS 118-46 and 118-82 had higher yields than both the average of the
parents and the yield of the better yielding parent. Three RILs, 118-94, 118-33, and 118-
39, were common to the goup with the 10 highest yielding RILs in three out of the six
environments (Table 10).

In Population 2, the mean yield of the 10 highest yielding RILs was higher than
the mean yield of the parents in all environments (Table 11). However, when the
individual yields of MCM and CELRK were considered, the mean yield of the 10 highest
yielding RILs was higher than the yields of both parents in only four of the six
environments (Table 11). When the high yielding RILs were compared to the check
varieties, the mean yield of the RILs was higher than that of the check varieties in all
environments (Table 11). For example, in Mich-1999, the 10 highest yielding RILs had a
mean yield of4145 ltg-ha", compared to the means of MCM and CELRK (3388 kglta"),
all the RILs (3459 kg-ha") and the check varieties (3582 ltgha") (Table 11). Unlike the
yield, the mean seed weight of these 10 high yielding RILs was not consistently higher
than the mean seed weights of the parents and the check varieties (Table 12).

Several RILs in Population 2 were common to the goup of the 10 highest yielding
RILs in more than two environments (Table 11). Some of these RILs had high yields
only in Michigan while some were high yielding in different Sites in different years. In
most cases, the yields of the RILs were comparable to or exceeded that of either parent.
Two RILs, one a dark red and another a light red kidney bean line, were among the 10

highest yielding RILs in four environments. These two RILs, 119-21 and 119-32, were

28

among the highest yielding entries in Mich-1998, Mich-1999, Minn-1996 and in NDak-
1999. Five RILs, four of which were light red kidney bean lines and one of which was of
a non-commercial seed color, were among the 10 highest yielding RILs in three
environments. One of these five, RIL 119-17, a light red kidney bean line, was in the
goup in three years in Michigan. The three other light red kidney bean RILs, 119-50,
119-70 and 119-79, were in the goup in three different environments. The yields and
seed weights of these and the rest of the RILs of the two populations, and of the check
cultivars in each experiment, are shown in Appendix Tables A.1 to A.4. In each
experiment, several RILs had yields higher than one or more of the commercial cultivars
used as checks.
Heritability Estimates and Correlations Between Yield and Seed weight

The ANOVA tables from which the variance components were estimated from
mean squares for yield and seed weight are shown in Tables A9 to A.12. Heritability
estimates for yield and seed weight were obtained using data fiom the 75 and 73 RILs,

respectively, of Populations 1 and 2. Estimates were moderate in value (Table 13).

Table 13. Heritability estimates for yield and seed weight of the 75 and 73 RILs in
Populations 1 and 2, respectively, calculated from data combined over Six
environments.

 

 

Population Yield8 (0") Seed Weighta (CIb)
Population 1 0.62 (0.45 - 0.71) 0.58 (0.38 - 0.68)
Population 2 0.63 (0.45 - 0.73) 0.69 (0.55 - 0.78)

 

' - Two replications in six environments; year-location combinations treated as
environments.
b CI — 95% confidence interval

29

Heritability estimates from Population 2 were higher than the values from Population 1.
In Population 1, the heritability estimates for yield and seed weight were 0.55 and 0.58,
respectively. In Population 2, the heritability estimates were 0.63 and 0.69 for yield and
seed weight, respectively (Table 13).

Seed weight was positively correlated with yield in both populations (Table 14).
In Population 1, the correlation coefficients ranged fi'om 0.4 to 0.7 in four environments -
Minn-1996, NDak-1999, Mich-1997 and Mich-1998). In Population 2, the coefficients of
correlation ranged from 0.2 to 0.6 in five environments - Mich-1996, Mich-1997, Mich-
1999, Minn-1996, and NDak-1999. For Population 2, seed color was also correlated with
yield and seed weight (Table 14). Numerical values for seed color (1 — light red; 2 —

mixture of light and dark red; 3 - dark red) were negatively correlated with

Table 14. Siglificant correlations between yield, seed weight and seed color in
Populations 1 and 2, planted in Michigan, Minnesota and North Dakota from 1996
to 1999. .

 

Environment'
Traitl Trait 2 Mich Minn Mich Mich Mich NDak Rangeb
1996 1996 1997 1998 1999 1999

 

Population 1

yield seed weight * "‘ * * * 0.42 to 0.66
Populatipn 2

yield seed weight * * * * * 0.17 to 0.60
yield seed color * * -0.21
seed weight seed color * * * * -0.26 to -0.45

 

' * - Siglificant at level of siglificance = 0.05.
b Range - Range of siglificant coefficients of correlation over environments.

30

yield in two environments (Mich-1996 and NDak-1999) with sigrificant coefficients of
correlation around —0.2. Thus, in these two environments, the light red kidney bean RILs
(seed color = 1) generally had Siglificantly higher yields than the dark red RILs (seed
color = 3). In Mich-1996, the light red kidney bean lines (35 RILs) had a mean yield of
3441 kg.ha", while the dark red kidney bean lines (27 RILs) had a mean yield of 3201
kg.ha'l (data not shown). In NDak-1999, the light red and dark red kidney bean lines had
mean yields of 1691 kg-ha’I and 1221 kg-ha", respectively (data not shown). The light
red kidney bean lines also had a higher mean yield overall (averaged over all
environments) (2686 ltg-ha") than the dark red lines (2518 lrgha").

Sigrificant coefficients of correlation between seed color and seed weight ranged
from —0.3 to —O.5 in four environments (Table 14). The negative correlations indicate
that in Mich-1996, Mich-1999, Minn-1996 and NDak-1999, the light red kidney bean
RILs had Sigrificantly higher seed weights than the RILs with dark red seed color.
Averaged over all the environments, the light red kidney bean lines had a mean seed
weight of 61 .1 g-100 seed", while the dark red kidney bean lines had a mean seed weight

of 58.2 g°100 seed" (data not shown).

31

DISCUSSION

Quantitative traits such as yield and seed weight are generally controlled by many
genes. The loci involved in the expression of a quantitative trait are called quantitative
trait loci. The effective manipulation of QTLS iS required for the improvement of the
traits they control. However, the individual effects of these QTLS are not readily
identifiable since the environment influences QTL expression to a Sigrificant but often
unknown degee. Sigrificant environmental effects also lead to genotype x environment
interactions, which obscure genetic variation. The environment affects not only the level
of performance of the genotypes, but also the degee of variation expressed in a
population as a whole. Thus, the reliability of cultivar performance across locations and
years is an important consideration in plant breeding (Fehr, 1987). If the genotype x
environmental interaction is substantial for a trait of interest, the breeder may have to test
over a series of locations and for several years to assess the breeding value of genotypes

under selection.

Some lines intended for commercial release perform well under a range of
locations and over several seasons while other lines are more limited in performance.
Information about a line’s performance in a series of environments is used to determine
its stability. Phenotypically stable genotypes are well buffered in the genetic sense and
show a predictable response to different environmental conditions. Stability is
particularly important for yield and yield components in dry bean (Kelly et al., 1998). To
ascertain the stability of a given set of materials, yield testing must be replicated over a

broad range of environments, including locations and years.

32

Genotype x environment interactions involving year effects warrant different
considerations in the breeding sense than do those interactions containing location terms
(Allard and Bradshaw, 1964). Genotype x year interactions generally are more
unpredictable than genotype x location interactions. The breeder has little control over
seasonal variations in rainfall, temperature, and cloud cover; however, environmental
influences due to location effects and genotype x location interactions may be
ameliorated by soil and crop management changes. Nevertheless, weather patterns and
disease incidence differ across locations too.

The results of the present study indicated that testing of beans for yield and yield
components (seed weight) for a period of years is necessary. In this study, the two
kidney bean recombinant inbred populations were evaluated over four seasons. The
variation attributed to sigrificant year effects may be more precisely determined from a
series of annual experiments such as was the case for the Michigan tests (four
consecutive years) than fi'om seasonal effects evaluated in a few randomly chosen
seasons (e.g., two years). Testing should thus be conducted over several consecutive
years to establish a genotype or goup of genotypes’ stability. Dry bean is extremely
responsive to high temperature, large diurnal fluctuations in temperature, drought, etc.
Testing in a limited number of seasons that are randomly chosen flour a seasonal interval
may preclude the breeder from accurately predicting a genotype’s stability for a trait.
Evaluation over several seasons will also allow a more precise estimate of the amount of
variation available for selection.

Evaluation of lines in more than one location allows the assessment of their

adaptation to different sites. The structure of the current study was such that location

33

effects cannot be determined for more than one year. In the two years in which the
populations were evaluated in more than one location, the locations involved were
Michigan and Minnesota in 1996, and Michigan and North Dakota in 1999. The
experiments involving two locations may be compared only within the year in which they
were conducted, using the results of Analyses 1 and 3 (Tables 4, 5, 8 and 9; Figures 1 and
2). In 1996, yield was Significantly higher in Minnesota than in Michigan for Population
1, but for Population 2, yield was sigrificantly higher in Michigan. In 1999, yield in
Michigan was sigrificantly higher than in North Dakota for the two populations.

Seed weight was Sigrificantly higher in Michigan than in either Minnesota or
North Dakota in 1996 and in 1999. In other yield trials, the yield of dry beans in
Michigan has been consistently higher than in either Minnesota or North Dakota
(unpublished data from cooperative dry bean nurseries fi'om 1994 to 1999). Comparisons
across different locations must be conducted over several years to obtain an accurate
assessment of location effects on yield and seed weight of bean genotypes. In this study,
Michigan can be compared with Minnesota or North Dakota in only one year. Thus,
other than the observations already given, no conclusions can be made about variable
yield and seed weight responses in the three locations, or about the plant characteristics
and‘developmental aspects that could account for these differences.

In the two populations tested in this study, no Single RIL was superior yielding
and manifested a high seed weight in all environments. Instead, ten RILs, which had the
highest yields in each environment, were identified. Although the mean yield of each
population (all RILs) was not higher than the mean of the respective parents, the mean

yield of these 10 RILs was higher than the means of both the parents and the checks.

34

Transgessive segegation of genes for yield might have contributed to these RILs
outyielding their respective parents. These results suggest that yield in kidney bean can
be increased by crossing established cultivars among themselves or cultivars by the
breeding lines. However, since the increases in yield were small, other sources of genes
for yield need to be introduced.

Some RILs were high yielding in at least two years in Michigan only. The
development of these RILs specifically for Michigan may be the appropriate and practical
approach. Some lines in both populations were among the highest yielding RILs in more
than one location, with yields higher than the parents and checks. Two RILs in
Population 1 (118-82 and 118-46) and two in Population 2 (1 19-18 and 119-32) were
common to the goup with the 10 highest yields in four of the six environments in which
the study was evaluated (Tables 10 and 11). Thus, the sigrificant effects of the
environment did not affect the ranking of some genotypes. At least some of RILs in
these kidney bean populations are apparently sufficiently stable across environments in
yield and seed weight. These results are not surprising since the three parents, MCM,
CDRK 82, and CELRK, had acceptable yields in these locations in previous yield trials
(unpublished data from cooperative dry bean nurseries fi'om 1994 to 1999), suggesting
that it may be possible to select particular genotypes that will perform well in all three
locations.

In addition to the statistical treatment of the environment as fixed effects, the
presence of significant interactions places a condition on inferences that can be made
about the main effects of genotypes, years, and year-location combinations

(environments). The estimates of these main effects are conditional, such that the

35

genotypic effects that may be concluded are only as observed in the years and
environments where the tests were conducted, and not over all possible enviromnents
(Freeman, 1973). Given Similar climatic conditions in future years of testing, the
performance of the RILs in the three locations, Michigan, Minnesota and North Dakota,
from 1996 to 1999 may be used only as a benchmark for potential yield. Environmental
fluctuations not sampled in these four years may cause results dissimilar to those reported
here. Likewise, the RILs may perform differently in areas other than these three
locations; i.e., no conclusions or predictions can be made about their yield potential in
other production areas.

Heritability estimates for yield and seed weight were mid-value for Population 1
and mid- to high-value for Population 2 (Table 13). These were similar to those reported
by Singh et a1. (1991) for a goup of mostly Andean genotypes. However, the variances
due to year, location and year x location interactions were confounded in the present
study, thus possibly causing an upward bias in the heritability estimates (F ehr, 1987).
Although these estimates aid in understanding the genetic control of these traits in kidney
beans, the very nature of heritability makes it clear that any estimate is specific both to
the material under study and to the structure of the experiment (Simmonds, 1979). The
heritability estimates from this study, along with the observed stability and yield potential
of some of the RILs, indicate yield in kidney bean can be increased through breeding and
selection. The high yielding RILs reported here may be used as parents in developing
lines with high yields and stable performance over several seasons. Lines that performed

well in Michigan over several years may be further developed specifically for the state

36

while those lines that had high yields in more than one location Showed a wider
adaptation.

Yield and seed weight were positively correlated in at least four environments in
the two populations. The correlation was in contrast to the findings reported by Nienhuis
and Singh (1985), who found negative correlations between the two traits. The
relationship between yield and seed weight is particularly important in dry bean, due to
the strict seed Size requirements placed on each market class. Kidney bean cultivars must
have seed size acceptable to the processing industry. The standard seed Size for this
market class is 50-65 gn per 100 seeds (Adams and Bedford, 1975). Beans that are
perceived as too small are undesirable by both producers and consumers. Breeders and
farmers, on the other hand, desire high yields. Thus, the positive correlations observed
here for the two kidney bean populations bode well for both bean breeders and
processors. The requirements of consistently high yielding lines and sufficiently large
seeds may be met without compromising one or the other.

Based on previous work by other authors (Adams, 1967; Nienhuis and Singh,
1985, 1988; Ranalli et al., 1991), the possibility that the correlated increases in yield and
seed weight were accompanied by compensatory reductions in number of seeds per pod
and/or nrunber of pods per plant exists. These relationships between yield components
are developmental in nature, are influenced by the environment, and may be due to
competition among plant structures for a common and limited nutrient supply (Adams,
1967). Environmental fluctuations may have triggered these mechanisms in the kidney
bean RILs used in this study. Further research is necessary to test this hypothesis. Low

but positive correlations between yield components were reported by Coyne (1968), who

37

 

 

also suggested the feasibility of selecting for one trait without an accompanying reduction
in the others. In the present study, since data on number of seeds per pod and number of
pods per plant were not taken, yield component compensation involving these two traits,
as they relate to yield and seed weight in kidney beans, warrant no further discussion.

In dry bean, several factors — lack of favorable alleles, low heritability, high
genotype x environment interactions, yield component compensation, low or negative
GCA within and between gene pools, and reliance on visual selection in early generations
contribute to the Slow progess in yield improvement (Kelly et al., 1998). The results
fi'om the present study with RILs from two kidney bean populations underscore the
influence of the environment on the expression of QTL controlling yield. Moderate
heritabilities for yield and seed weight indicate sufficient genetic control over the trait to
permit successful breeding for increased yield.

The major limitation in yield breeding in dry bean is not low heritability, stability
or genotype x environment interactions, but a lack of favorable genes for yield in the
current cultivated gerrnplasm (Kelly et al., 1999). Since kidney beans have a narrow
genetic base, new sources of genetic material are necessary to introduce new genes for
yield into existing gerrnplasm pools. There is a need to identify and utilize favorable
genes from other sources such as plant introductions and wild accessions of P. vulgaris.
Kelly et a1. (1999) proposed a three-tiered approach to yield breeding, which utilizes a
broad genetic base as a source of genes for elite lines. Such approaches will take
advantage of the diversity of bean gennplasm and ensure continued success in increasing
the yield of kidney beans and other classes of dry bean. Utilizing new sources of genes

for yield, however, may have undesirable effects on other traits considered important for

38

commercial kidney bean cultivars, such as canning quality. Unadapted germplasm may
have the necessary genes to increase yield in cultivated genotypes but have not been
subjected to selection for traits such as wholeness of beans after processing and general
acceptability for consumption. Thus, the introduction of genes from these unadapted
sources may compromise canning quality. Such negative correlations, if present, retard
progess in breeding for yield (Yan and Wallace, 1995). Both sets of traits must be
evaluated and monitored throughout the breeding process in order to meet the desired

goals for yield breeding without compromising other important traits.

39

CHAPTER 2: EVALUATION OF CANNING QUALITY IN KIDNEY BEAN,
AND THE IDENTIFICATION OF RANDOM AMPLIFIED POLYMORPHIC
DNA (RAPD) MARKERS ASSOCIATED WITH CAN NING QUALITY TRAITS

INTRODUCTION

Much of the dry bean production in the US. that is canned commercially is
consruned domestically. Due to their nutritional composition, dry bean is a valuable
addition to the diet of consumers. Kidney beans, for example, are composed of
approximately 22% protein, 4.0% ash, 67% carbohydrates and 7.0% fiber (Sathe et al.,
1984). In addition, beans have a long Shelf life and cost less than most animal, fi'uit and
vegetable products.

Since dry beans are eaten as whole gains and not milled into flour, consumers
have been conditioned by years of use to expect certain characteristics of the dry, soaked,
and cooked beans. Likewise, processors have their own set of criteria that are mostly
concerned with processing efficiency and profitability. Due to consumer expectations
and processing standards for beans, the dry bean processing industry has made processing
characteristics a major consideration in their choice of bean varieties. Plant breeders and
food scientists collaborate to ensure that newly released bean varieties meet, not only
yield expectations, but also the acceptability standards established by the processing
industry for the various market classes of dry bean.

In view of the steps necessary to prepare beans for eating, a priori tests that
evaluate components of canning quality have been developed (Hosfield and Uebersax,
1980; Hosfield et al., 1984a; Ghaderi etal., 1984; and Walters et al., 1997). These tests

measure distinct physical and chemical properties of bean seeds that are logically related

40

to canning quality (Hosfield et al., 1984b). Several canning tests have been adopted for
use in private and public breeding progams. The test measurements do not fall into
discrete measurement classes and hence, are quantitative in nature. Moreover, canning
quality methodology in dry bean generally requires the use of advanced generation plant
material to ensure sufficient seed for evaluation.

Breeding for canning quality in dry bean provides a difficult challenge to the plant
breeder because of the quantitative nature of the component traits, and the necessity of
waiting until the F5 or F6 generation when sufficient seed is available for canning tests.
Indirect selection using linked molecular markers, termed marker-assisted selection
(MAS), has received attention as a method for increasing selection efficiency within
breeding progams. If a trait is expensive to evaluate, under polygenic control, or
considerably influenced by the environment, MAS may be more efficient than traditional
(direct) selection methods based on phenotype. The use of MAS has proven to be
effective in shortening the time involved in the improvement of quantitative traits in
many crops (Dudley, 1993), and may prove usefirl in breeding for canning quality in
beans, in general, and kidney beans, in particular.

Kidney beans constitute a sigrificant percentage of dry bean production in the
US. Light red kidney beans are used in chili products, while the dark red varieties are a
sigrificant component of restaurant salad bars. A large portion of the annual kidney bean
crop in the US. is canned prior to commercial distribution, and thus must meet the
standards required by the bean canning industry and by consumers. Canning quality thus
continues to be an important focus for kidney bean breeding progams. In addition to

conventional approaches, improved technology, such as the development of molecular

41

markers for complex traits, has afforded the use of methods not previously available to
plant breeders. The future of plant breeding includes the assessment of the feasibility of
using these methods and their effective application to problems with which breeders have
been dealing for decades. This present study seeks to address and remedy the lack of
information on RAPD markers associated with canning quality traits in beans in general
and kidney beans in particular.

The study was conducted on two recombinant inbred populations of kidney bean.
The populations were planted in Michigan fi'om 1996 to 1999, in Mirmesota in 1996, and
in North Dakota in 1999. The objectives of the research were to a) evaluate canning
quality of the two recombinant inbred lines in six environments; b) estimate heritabilities
and correlations between canning quality traits; 0) identify putative RAPD markers for
canning quality; and (1) determine whether markers associated with canning quality in

navy bean are useful for kidney bean.

REVIEW OF LITERATURE

The importance of canned beans in the diets of many people has prompted studies
dealing with the various components of canning quality and the development of methods
for evaluating these components (Hosfield and Uebersax, 1980; Hosfield et al., 1984a;
Ghaderi et al., 1984). Advances in biological research, such as the use of molecular
markers, have opened the possibility that such markers may facilitate canning quality
evaluation and lead to the development of varieties that meet the requirements of

processors and consumers (Walters et al., 1997).

42

.11.?

 

 

Canning Quality

Canning quality is composed of traits that affect the hydration characteristics of
swds, thermal conditions that render the seed palatable and provide for the digestion of
nutrients, and consumer expectations for the cooked product. Some traits that processors
and consumers pay attention to are: rate of water uptake, volume increase of seeds,
expansion coefficients of soaked and blanched seeds, brine characteristics, uniformity of
seed Size and shape, seed color and appearance, mouthfeel, texture, digestibility, degee
of clumping and splits, visual appeal (perceived overall acceptability), net weight after
canning (processors’ yield), flavor, and ease of preparation and cooking (Adams and
Bedford, 1975; Hosfield and Uebersax, 1980, 1990; Uebersax and Bedford, 1980;
Ghaderi et al., 1984; Hosfield et al., 1984a; Hosfield, 1991; Fomey et al., 1990; and
Walters, 1995). Although these physical and chemical attributes of cooked beans all
contribute to the definition of processing quality, no single trait defines overall
acceptability (Hosfield and Uebersax, 1990; Hosfield, 1991).
Commnents of canning Quality. Rapid and uniform uptake of water during soaking is a
desirable trait of beans for canning (Hosfield and Uebersax, 1990; Adams and Bedford,
1975). A moisture content of 55% after soaking is considered optimum (Uebersax,
1985). Soakability is generally measured as the difference in weight of a bean sample
before and after soaking, and is expressed as the hydration coefficient (HC) (Adams and
Bedford, 1975).

Texture (TXT) is another primary canning quality character. TXT affects the
perceived stimulus for chewing, and hence, influences to a large degee a consumer’s

acceptance of a food product. TXT of processed beans has three components: firmness,

43

gumminess, and adhesiveness. Firrnness is defined as the resistance of a bean to
deformation after a mechanical force is applied. Lu and Chang (1996) and Van Buren et
a1. (1986) reported contrasting effects of firmness on the degee of Splitting of cooked
beans. According to Van Buren et al. (1986), a low incidence of splitting is associated
with a lower WDWT and firmer cooked beans. Lu and Chang (1996), on the other hand,
reported that high firmness values are associated with a more viscous medium after
cooking and more splits, and thus contribute to a lower overall acceptability (visual
appeal) of the cooked beans. Harvest date appears to affect the firrrmess of processed
beans, with later dates resulting in firmer textures (Kays et al., 1980). Gumminess is
measured by the energy required to disintegate the sample and adhesiveness is the
degee of stickiness or difficulty of removing the substance from a smooth surface, e.g.,
the roof of the mouth. A panel of judges who render an opinion of a perceived stimulus
may subjectively evaluate TXT. TXT can also be estimated objectively by using an Allo-
Kramer Shear Press (Food Technology Corp., Rockville, MD). Although the firmness of
a food, as determined with a shear press, ignores other perceptions, such as viscosity of
the medium, adhesion, or gumminess, it estimates TXT in a practical sense. AS such, the
measurement serves as an index for consumer acceptance. In the case of cooked beans,
beans may be unacceptable if perceived as too firm (“tough beans”) or too soft (“mushy
beans”) (Hosfield et al., 1984a).

The hydration properties of cooked beans are expressed as the washed drained
weight (WDWT), which is the net weight of processed beans after rinsing under cold tap
water and draining (Wassimi et al., 1990; Hosfield and Uebersax, 1980). This trait is

important to canners who seek high can yields because, when the WDWT is high, fewer

beans are required to fill a can of a particular volume. Beans with a high WDWT usually
have high swelling capacities and high physical entrainment brought about by water-
macromolecule interactions. A low WDWT may occur when beans lose excessive solids
during processing. With excessive solids loss, water entrainment is low, and low WDWT
occurs because the solids lost are heavier than the water absorbed. In general, WDWTS
for canned beans with initial fresh weights equivalent to 100g total solids (TS) range
from 275-375 g. Higher WDWTS have been associated with softer beans after canning
(Lu and Chang, 1996).

The degee of clumping and splitting are physico-chemical attributes of cooked
beans that have a marked influence on visual appeal, which is one of the primary criteria
of consumers of beans. Clumping may be due to excessive starch exudation during
canning (Adams and Bedford, 1975) and is undesirable (Wang et al., 1988). Fewer splits
in canned beans contribute to higher acceptability (Lu and Chang, 1996; F omey et al.,
1990). Splitting appears to be affected by seed Size, with larger seeds showing more
splits (F omey et al., 1990). Later harvest dates seem to result in fewer split seeds (Kays
et al., 1980). Splitting may also be affected by threshing and post-harvest handling
conditions, as well as soaking and processing conditions.

Brine characteristics after processing are also important for beans processed in tin
cans or glass jars. Consistency, gaininess or cloudiness, and color of the brine are
considered in rating brine characteristics for acceptability. Brine of good quality is
slightly viscous, clear, without obvious starch ganules, and drains easily from the whole

beans (Adams and Bedford, 1975). Brine that is highly viscous has been correlated with

45

geater clumping than less viscous brine. Correlations have shown that the more starch in
the brine, the lower the overall acceptability of beans (Lu and Chang, 1996).

Other factors affecting overall visual appearance are uniformity of seed size and
shape in a sample, the intensity and uniformity of seed color, wholeness of the beans, and
absence of loose seed coats and other extraneous material (Adams and Bedford, 1975;
Ghaderi et al., 1984; and Fomey et al., 1990). Although most of the qualities discussed
above are based on sensory perception, certain procedures have been established in order
to objectively evaluate each component trait (Hosfleld and Uebersax, 1990).

Procedures in processing. Canning methods employed during genotype evaluation
should Simulate those used in the commercial canning industry, with the primary purpose
being aroma development and rendering the beans tender enough for human
consumption. Processing also removes or inactivates beany or bitter flavors and
antinutritional factors such as protease inhibitors, lectins, phenolic compounds and
phytates (Deshpande et al., 1984).

According to Adams and Bedford (197 5), the procedures appropriate for the
evaluation of canning quality are as follows: selection of good quality raw dry beans,
equilibration of moisture content, soaking, blanching, filling in cans, cooking,
equalization of cooked beans and evaluation. In addition to bean genotype and moisture
level, the different conditions produced by these procedures affect the quality of canned
beans.

Uebersax (1972) studied the effects of storage and soaking methods on the
processing quality of navy beans. The temperature and relative humidity under which

beans are stored were found to affect processed bean color, flavor and firmness

46

(Uebersax, 1972). In the same study, the temperature and composition of the soak water
was found to Sigrificantly affect water uptake, bean volume and texture. Further research
(Uebersax and Bedford, 1980; Wiese and Jackson, 1993) corroborated these results.

If processing evaluations are to measure true differences between varieties, the
moisture content of each seed sample Should be equilibrated to a common value of about
14 to 18% (Deshpande et al., 1984). Adams and Bedford (1975) suggested moisture
levels of 12-14%, if the beans are to hydrate and cook readily. Beans stored under high
RH, which would consequently have high moisture contents, require a longer cooking
time (K011 and Sanshuck, 1981). On the other hand, Deshpande et a1. (1984) observed
that, if the moisture content is too low, the beans may not imbibe water normally and
become hard to cook, or the seed coats may become brittle and crack during processing.

Soaking ensures tenderness and uniform expansion of the beans during canning
(Hoff and Nelson, 1965), shortens the processing time, and reduces the amount of toxic
compounds found in raw beans (Deshpande et al., 1984; Uebersax et al., 1991). Van
Buren et a1. (1986) reported that higher concentrations of calcium in the soaking medium
(150-350 ppm) and higher soak temperatures (66-71°C) Sigrificantly reduce splitting.

Uebersax and Bedford (1980) determined that the following two-step process
provided optimum soaking conditions for canning beans: 30 minutes at 23°C and 30
minutes at 88°C with at least 50 ppm calcium in the soak water. Addition of 100 ppm
calcium ion resulted in beans with minimum damage due to splitting and beans becoming
mushy (Hosfield and Uebersax, 1980). Beans soaked at 82°C or 93°C for 30 minutes had
short rehydration times and hydration coefficients similar to beans soaked in many

processing plants (ngal and Davis, 1994). In the same study, processing conditions at

47

121°C for 21 minutes were found to result in higher WDWTS, softer beans, and less
splitting than the control, which was processed at 116°C for 41 minutes.

Blanching eliminates air and equalizes moisture in the samples. However,
overblanching beans causes the seed coats to split (Adams and Bedford, 1975). Steam
blanching at a high temperature for a short time produced canned beans with good
quality, although quality varied with cultivar and length of time of the blanching process
(Drake and Kinman, 1984).

Addition of both calcium chloride (CaClz) and ethylenediaminetetraacetic acid
(EDTA) to the processing medium improved both the firmness and color of processed
beans, and resulted in less splitting in kidney beans (Van Buren, 1986). The use of CaClz
alone reduced clumping and splitting of beans (Wang et al., 1988; Wang and Chang,
1988). Shorter cooking times also reduced splitting in kidney beans (Van Buren, 1986).
After beans are processed, they continue to imbibe water until they reach a moisture
content of approximately 65% (Adams and Bedford, 1975). Storing processed beans for
two weeks before evaluation ensured that water imbibition in the can was complete
(Hosfield and Uebersax, 1980).

In addition to storage conditions and processing procedures, processing quality in
dry'bean depends on the genotype, environment, their interactions, and the condition of
seeds at harvest (Wassimi et al., 1990; Hosfield, 1991; Lu and Chang, 1996; Nordstrom
and Sistrunk, 1979; Junek et al., 1980; Hosfield et al., 1984b). In studies by Uebersax
and Bedford (1980), Ghaderi et a1. (1984), Hosfield et al. (1984b), Wassimi et a1. (1990)
and Walters et a1. (1997), environmental effects on certain processing quality traits were

found to be Siglificant. The variations in the phenotype caused by the environment are

48

usually unpredictable. The responses of different genotypes relative to one another may
vary over sites and years, and fiequently lead to genotype by environment interactions,
which complicate the interpretation of results (Hosfield, 1991). Sigrificant interactions
between genotype and environment must be considered in interpreting the effect of
genotype or environment alone.

Genetics of Canning Qpality. Genetic variation with respect to processing quality has
been reported in dry beans (Hosfield and Uebersax, 1980, 1990; Hosfield et al., 1984b;
Wassimi et al., 1990). AS for any complex trait, the breeder must have knowledge of the
genetic control and heritability of the traits comprising canning quality in order to
ascertain and utilize phenotypic variability for the traits under selection.

Wassimi et a1. (1990) confirmed the mode of inheritance of physico-chemical
traits related to processing quality in dry bean. Genes that behaved in an additive fashion
predominated over non-additive ones for soaked bean weight (SBWT), soaked bean
water content (SBWC), splitting (SPLT), and the washed-drained weight coefficient
(WDWTR). Clumping (CLMP), WDWT and TXT were influenced by genes that
behaved in both an additive and a non-additive fashion. In the same study, most genes
for WDWT and TXT were found to be completely dominant. Heritability estimates
obtained by Walters et a1. (1997) were moderate to high: 0.59 for visual appeal (VIS),
0.64 for TXT and 0.67 for WDWT.

Correlations exist among the various parameters of canning quality. Hosfield and
Uebersax (1980) reported that soaking properties were not correlated with textural
differences among the tropically adapted genotypes included in their study. However,

Ghaderi (1984) reported a negative correlation between TXT and WDWT (a hydration

49

property) of cooked beans. Hosfield et al. (1984a) looked into the question of trait
interrelationships in black seeded dry bean in a multivariate analysis of processing
quality. Factor analysis (Catell 1965 a, b; Kim, 1975) indicated that soaking, cooked
color, thermal and dry color traits were orthogonal although TXT and WDWT, two
thermal traits, were negatively correlated (Hosfield, et al., 1984a). In three populations of
navy bean, Walters et al. (1997) detected negative correlations between TXT and WDM
(r = - 0.53 to - 0.83), and between VIS and WDM (r = - 0.26 to - 0.66). In the
terminology used by Walters et a1. (1997), VIS was visual appeal, a perception of the
overall appearance of canned beans, and WDM was the washed drained mass, equivalent
to WDWT. The same authors (Walters et al., 1997) reported sigrificant and positive
correlations between VIS and TXT (r = 0.19 to 0.66). Since the correlations are
phenotypic in nature, they may be due to the combined effects of genotype and the
processing environment, and do not necessarily reflect associations due to genetic factors
such as linkage or pleiotropic effects (N ienhuis and Singh, 1985).
Use of Markers in Crop Improvement

A quantitative trait is more difficult to improve than a Mendelian character
because the type and degee of influence of several loci acting in concert on a particular
trait cannot be identified easily (Dudley, 1993 ), unlike for a single-gene trait where each
allele results in a distinct phenotype. The number of genes involved and the interactions
among them imply that several loci must be manipulated at the same time to obtain the
desired phenotype (Ribaut and Hoisington, 1998).

Canning quality in dry bean is viewed as a “super trait” because no single variable

can adequately describe the properties preferred in and required of a sample (Hosfield

50

and Uebersax, 1990). In view of the inherent complexity, a “super trait” is difficult to
improve. At best, the breeder seeks to dissect them into a number of component
characters that can be individually measured and selected (Hosfield and Uebersax, 1990).
In improving the processing quality of dry beans, the breeder must consider that additive
and/or dominance effects influence each of the component traits separately. The effect of
the environment on the expression of each component trait must also be taken into
account. Thus, the evaluation of a large number of samples with small differences using
objective and/or subjective methods would be difficult (Ghaderi et al., 1984).
Furthermore, for traits that have low heritabilities and high additive variance, selection
using conventional methods should be done in later generations, such as the F6
generation, when the lines are nearly homozygous (Elia et al., 1997). Technological
advances in the last decade have given plant breeders an impetus to reevaluate the use of
genetic markers to address these problems in various crops.

For simply-inherited traits, marker-assisted selection (MAS) has been used to
select in early generations and to reduce the size of the population used during selection
(Staub and Serquen, 1996). MAS is of particular value in breeding for characters with
low heritabilities and when the marker is associated with additive genetic variance (Staub
et al., 1996). In quantitative trait analysis and breeding, the use of markers and genetic
maps has permitted the identification of regions of the genome that most likely contain
the genes or goups of genes [quantitative trait loci (QTL)], responsible for the expression
of these traits. Molecular markers may also aid in understanding genotype x environment
interactions when Siglificant marker-QTL associations are compared in different

environments (Dudley, 1993). Markers also allow the comparison of the genomes of

51

different but related taxa with regards to the location of common QTL (Paterson, 1995).
By knowing the locations of important QTL in the genome, one can facilitate their
precise manipulation in breeding progams. However, even if the exact locations of QTL
in the genome are not known, the associations of the genes with easily identifiable
markers may aid in trait improvement by combining MAS with conventional breeding.
MAS is an indirect selection method that appeals to breeders because it enables
them to select in early generations, which can reduce both the time and the cost of the
selection process. Eathington et a1. (1997) used marker-QTL associations to predict the
yield performance of maize in later generations of testcrosses using data fiom earlier
generations. Knapp (1998) proposed the use of MAS to increase the probability of
selecting superior genotypes and predicted that, for traits with low to moderate
heritabilities, MAS will require fewer resources to reach a selection goal when the
selection intensity is high. Markers have also been used successfully to improve disease
and insect pest resistance, and other characteristics of crop species (Haley et al., 1993;
Young and Kelly, 1996; Kelly and Miklas, 1998; Kelly and Miklas, 1999). But before
MAS can be used in a breeding progam, associations between appropriate marker alleles
and QTL must be identified.
Mogphological and Protein Markers. The first markers reported were easily observed
phenotypic characteristics associated with economically important traits. Associations of
simply inherited traits (markers) with more complex characteristics were reported as
early as 1923 when Sax documented the association of seed size with a seed coat color
marker in P. vulgaris. Since then, numerous authors working with various crops have

found other associations between Simply inherited characters and quantitative traits.

52

The use of physical and chemical characteristics to predict dry bean canning
quality has been proposed by various authors. The pasting viscosity of whole bean flour
was highly correlated with the texture or firmness of canned navy beans in separate
studies (Ruengsakulrach, 1994; Lu etal., 1996). The authors of the two studies suggested
that pasting characteristics might be useful in screening breeding lines for canning quality
in early generations. Lu et a1. (1996) found correlations between pasting viscosity and
WDWT, and between viscosity of the canned bean medium and overall acceptability. Lu
et a1. (1996) also suggested using the hydration ability of raw navy beans to predict the
degee of color of cooked beans, and the turbidity of micro-cooked bean liquid to predict
the clarity of the canned bean medium. However, Ruengsakulrach et a1. (1994) suggested
that the color of cooked beans might be a function of processing time and the
cararnelization of sugars during heating, implying that physico-chemical processes such
as hydration will not have any effect on the final color of the cooked beans.

In kidney beans, Siglificant correlations have been reported between the relative
amount of damaged beans after processing and both bean density and seed coat weight
(Heil et al., 1992). These researchers (Heil et al., 1992) suggested that these physical
pr0perties could be used to estimate bean damage during processing, and in aiding dry
bean breeders in improving processing qualities. In addition, soluble pectin content was
highly correlated with firmness in various dry bean cultivars (Wang et al., 1988). TheSe
authors suggested that pectin content could thus be used as a parameter for screening
lines for desirable firmness of cooked beans.

Some disadvantages of using physical and chemical traits as markers are the

limited number available and undesirable phenotypes of many of these markers, and, in

53

the case of cytological markers, the large Size of the chromosomes and chromosome
segnents used (Dudley, 1993). Morphological markers also rely on recessive mutations
and require much time to develop (McClean et al., 1994). One improvement on the use
of morphological markers for selecting QTL was the employment of isozymes, which
have been used in MAS in several crops such as maize (Stuber and Edwards, 1986).
Although these marker systems have proved useful in genetic studies, their biochemical
nature and function limit the number of enzyme systems commonly used for analysis to
about 40-60 reactions (Gabriel, 1971; Gottlieb, 1982; Burow and Blake, 1998). The lack
of potential isozyme markers limits their use in QTL analysis, fine mapping, and MAS.
The same limitation exists for other protein marker systems. As a consequence,
molecular markers that function at the DNA level have largely replaced protein marker
analysis in gene-tagging experiments (Burow and Blake, 1998).
DNA Markers. DNA markers first became widely used in genetic analysis in the 1980s,
with the advantage of an increased number of potential markers available (Burow and
Blake, 1998). DNA markers have also proved to facilitate faster recovery of genomic
segnents, more efficient selection, and even the transfer of favorable alleles fi'om wild
relatives to elite cultivars (Ribaut and Hoisington, 1998).

There are several criteria to be met in selecting molecular markers for use in
MAS. The value of a molecular marker depends on its inherent repeatability, map
position and linkage with an economically important trait (Weeden et al., 1992; Staub
and Serquen, 1996). Linkage of 10 cM or less is helpful to increase gain fiom selection
(Paran et al., 1991; Kennard et al., 1994; Timmerrnan et al., 1994). Miklas et a1. (1995)

listed criteria necessary for a marker to be useful for indirect selection of quantitative

54

traits: 1) relative Stability across environments; 2) variation that accounts for as much as
or more than the heritability of the trait being considered; and 3) in the case of disease
resistance, presence only in the resistant germplasm.

The first DNA markers used in QTL analysis were restriction fiagnent length
polymorphisms (RF LPs). Labeled probes detect RFLPS as variable sized DNA fragnents
generated by restriction enzymes that cut the DNA at specific sites in the molecule.
RFLPS behave as codominant markers, viewed as bands on cellulose-acetate film (Staub
et al., 1996). They were first used to construct a human genetic map (Botstein et al.,
1980). Since the time of their pioneering use in human genetic studies, RFLPS have been
applied in the construction of genetic linkage maps in crops such as maize and tomato
(Helentjaris et al., 1986), and the tagging of QTL such as those controlling the amount of
soluble solids in tomato (Osborn et al., 1987). Other examples of DNA marker systems
are restriction landmark genome scanning (RLGS), rrricrosatellite systems, sequence-
tagged sites (STS) and amplified fragnent-length polymorphism (AFLP) (Burow and
Blake, 1998), and random amplified polymorphic DNA (RAPD).

A RAPD marker makes use of the polymerase chain reaction (PCR). PCR-based
markers require small amounts of DNA to be used as a template, and thus, allow early
sampling and rapid DNA preparation. Large sample sizes can also be handled efficiently
(Ribaut and Hoisington, 1998). A RAPD is generated through the amplification of
genomic DNA by using single primers usually 10 nucleotides long and of arbitrary
nucleotide sequence (Williams et al., 1990). Low stringency amplification and the short
lengths of the primers make possible multiple binding sites throughout the genome.

Amplification of DNA fiagnents, the sequences of which are unknown, occurs when two

55

binding sites are in close proximity (Burow and Blake, 1998). RAPDS usually behave as
dominant markers, scored as the absence or presence of a particular band, and are
inherited in a Mendelian fashion (Williams et al., 1990; Staub et al., 1996).
Polymorphisms may be due to mutations or deletions in the primer-binding Site,
insertions that increase the distance between binding sites or insertions that change the
size of a DNA segnent without preventing its amplification (Williams et al., 1990;
Burow and Blake, 1998).

In dry bean, RAPD markers are comparable to RFLPS with regard to the
frequencies of polymorphisms observed (Miklas and Kelly, 1992). However, RAPD
technology has the following advantages over RFLPS and other molecular marker
technology: 1) specific nucleotide sequence information of primers or clones is not
required to generate the polymorphisms; 2) a universal commercially available set of
primers can be used for genomic analysis; 3) preliminary work, such as cloning and
isolation of DNA probes and preparation of filters for hybridization, is not required; 4)
the method can be automated, which allows the running of large numbers of samples
simultaneously; 5) labor-intensive Southern blot hybridizations are not employed; and 6)
only small quantities of DNA are needed, allowmg the analysis of limited samples and
eliminating extraction of large amounts of DNA (Williams et al., 1990; Burow and Blake,
1998).

The use of RAPD markers also has its disadvantages. Marker patterns fiom
RAPD analysis are not as reliable or reproducible as those obtained fi'om RFLP analysis,
because of the low stringency of the amplification conditions used, variations in DNA

quality and concentration, and optimal primer concentrations (Burow and Blake, 1998).

56

As is the case for any marker system relying on dominant alleles for resolution, RAPD
markers cannot resolve heterozygous from homozygous loci (Williams et al., 1990).
When RAPDS are used, the sequence homology of similarly-Sized fragnents is
impossible to discern, unmapped markers are inefficient to use for genetic analysis, and
clustering of markers may occur in some instances (Burow and Blake, 1998).

RAPD markers have been used to construct and aligr linkage maps (McClean et
al., 1994 and Freyre et al., 1998), determine genetic relationships between genotypes
(Beebe et al., 1995; Skroch et al., 1992), and tag Single genes and QTL in several crop
species, including dry beans (Bai, 1996; Miklas et al., 1995, 1996, 1998a). Most of the
RAPD research in dry bean involves resistance to diseases, such as bean rust [caused by
the pathogen Uromyces appendiculatus (Pers.:Pers.) Unger] (Jung et al., 1996), bean
golden mosaic (BGM) (caused by a geminivirus) (Miklas et al., 1996), charcoal rot or
ashy stem blight [Macrophomina phaseolina (Tassi) Goidanich] (Olaya et al., 1996),
common bacterial blight (CBB) [Xanthomonas campestris pv. phaseoli (Smith) Dye]
(Jung et al., 1996), and anthracnose (Colletotrichum lindemuthianum [(Sacc. & Maglus)
Lams.-Scrib.]) (Melotto etal., 1998). Single genes control most of these resistance traits.
Kelly and Miklas (1999) provide a comprehensive review of the markers, which have
been identified for various resistance genes.

Jung et a1. (1996) identified RAPD markers for QTL governing plant architecture
and resistance to common blight and web blight [Thanatephorus cucumeris (A.B. Frank)
Donk], using a population of recombinant inbred lines of common bean. A partial
linkage map covering 545 centirnorgans (cM) and including 75 of 84 markers studied

was made. Miklas et a1. (1996) established linkages between RAPD markers and a QTL

57

conditioning BGMV or CBB resistance. This Study resulted in 14 RAPD markers being
selectively mapped in the population. Miklas et al. (1998b) analyzed QTL for field
resistance to ashy stem blight. Park et a1. (1998) identified and mapped RAPD markers
associated with QTL for seed Size and shape.

Several authors have also identified RAPD markers linked to genes controlling
various aspects of quality in crops. Examples of these traits are the milling energy
requirement of barley (Hordeum vulgare L.) (Chalmers et al., 1993), oleic acid
concentration in spring turnip rape (Brassica rapa L. ssp. oleifera DC.) (Tanhuanpiifi et
al., 1996) and fruit ripening in tomato (Lycopersicon esculentum L.) (Doganlar et al.,
2000). With regards to processing quality in beans, almost no work has been done using
molecular markers. In a study conducted over two years and in two locations, Walters
(1995) identified several RAPD markers associated with some component traits of
canning quality in three populations of navy bean. The populations were screened with
390 primers. Markers were linked to VIS, TXT and WDWT of processed beans. Results
of the study Showed location- and population- specificity among the marker-QTL
associations identified.

Selgtive Genotyping. Most QTL mapping experiments involve large populations,
usually composed of more than 200 individuals (Paterson, 1998). Several strategies have
been proposed to make use of smaller populations or to increase the efficiency of
handling large ones, without sacrificing the amount and quality of the information that
can be obtained. Some of these approaches are selective genotyping (Lander and
Botstein, 1989; Paterson, 1998) and DNA pooling strategies (Michelrnore et al., 1991;

Wang and Peterson, 1994; Darvasi and Soller, 1994; Paterson, 1998).

58

The use of bulked segegant analysis has Simplified genetic mapping by reducing
the number of lines, which are initially screened for putative markers (Michelrnore et al.,
1991). In this method, the individuals in each phenotypic extreme as a single DNA
sample or bulk (Paterson, 1998). Within each bulk, the individuals are presumably
identical for a trait or genomic region of interest but are arbitrary for the others
(Michelmore et al., 1991). Effectively, the two bulks differ only in the target region, and
are heterozygous and monomorphic for other loci. The goal is to identify markers that
distinguish the two bulks and thus presumably are linked to the target locus. These
markers differ between the bulks in their presence or absence, or in the intensity of the
bands observed, depending on their distance from the target locus. The putative markers
are then confirmed and mapped by genotyping the entire population. This approach
eliminates the need to initially screen the entire population with all possible markers.
Bulked segegant analysis was originally used for single-gene traits but was also
proposed to be useful for mapping QTL (Michelrnore et al., 1991). Examples of Simply
inherited characters for which markers have been found using this approach are nematode
(Heterodera schachtii Schm.) resistance in sugar beet (Beta vulgaris L.) (Heller et al.,
1996) and fruit skin color in apple (Malus sp.) (Cheng et al., 1996).

Selective genotyping is effective for QTL that affect only one phenotype
(Paterson, 1998). Using this method, a large population iS generated and evaluated
phenotypically, but genotyping is done only on those individuals that exhibit the most
extreme phenotypes (Lander and Botstein, 1989). Since phenotypic evaluation frequently
costs less than genotyping, it is more efficient to increase the number of progeny while

genotyping only a subset of individuals, than to genotype the entire population. Selective

59

genotyping has been used to identify markers linked to various QTL such as those
involved in disease resistance in common bean (Miklas et al., 1996) and tomato (Chagué
et al., 1997), oleic acid concentration in spring turnip rape (Brassica rapa L. ssp. oleifera
DC.) (Tanhuanpaa et al., 1996), and milling energy requirements in barley (Hordeum
vulgare L.) (Chalmers et al., 1993).

While DNA pooling methods are effective for rapidly identifying markers, several
authors have observed some limitations on their applicability to QTL mapping (Darvasi
and Soller, 1994; Wang and Paterson, 1994; Paterson, 1998). Although selective
genotyping methods are able to detect QTL with large effects, they may not detect the
majority of QTL that have small phenotypic effects (Darvasi and Soller, 1994; Wang and
Paterson, 1994). Other factors such as segegation distortion and dominance may also
influence the allelic composition of the DNA pools, resulting in false positive reactions
and complicating the utility of these approaches (Wang and Paterson, 1994; Paterson,
1998). Wang and Paterson (1994) suggested the following to reduce the occurrence of
false positives when using DNA pooling approaches: a) use parents with extreme
variation for the trait of interest, b) use large populations, 0) use homozygous populations
such as recombinant inbred or doubled haploid lines, c) and replicate the phenotypic

evaluations.

MATERIALS AND METHODS

The two recombinant inbred populations were derived fiom crosses between

‘Montcalm’ (MCM) and ‘California Dark Red Kidney 82’ (CDRK 82), and MCM and

60

‘California Early Light Red Kidney’ (CELRK). MCM (Figure 3), the common parent of
the two populations, has a long-standing reputation in the canning industry for its
desirable canning quality. Compared to MCM, CDRK 82 (Figure 4) and CELRK (Figure
5) have less appealing canning quality. The recombinant inbred lines (RILs), parents and
checks for the two populations were planted in separate experiments at the Montcalm
Research Farm near Stanton, Mich. in 1996, 1997, 1998 and 1999; in Perham, Minn. in
1996 and in Erie, NDak. in 1999. The details of planting were described in Study 1.
Evaluation of Canning Quality

After harvest, threshing and cleaning of the seeds, a digital moisture computer
(Burrows Model 700) was used to determine percentage moisture of 250g samples of the
seeds of each entry. Beans fi'om each field plot with a flesh weight equivalent of 100 g
total solids (Hosfield and Uebersax, 1980) were placed in nylon mesh bags and soaked at
21 C for 30 minutes and blanched at 88 C for 30 minutes. Two replicates of each bean
sample were processed. The cold soak and blanch were done in distilled water adjusted
to 100 mg-L’l calcium ion. The soaking procedure resulted in a sample with minimum
damage similar to beans soaked continuously in the high-temperature systems common
throughout the US. canning industry (Hosfield and Uebersax, 1980). The duplicate
samples of soaked and blanched beans from each field plot were placed into No. 303 (100
x 75 mm) tin cans and weighed. Soaked bean weight (SBWT), the weight (g) gained by
the beans through water imbibition during soaking and blanching, was obtained for each
replicate. The hydration coefficient was calculated as follows:

SBWT

 

HC =
fresh weight (Hosfield and Uebersax, 1980)

61

f Monfcal—ni

 

Figure 3. Processed beans of the cultivar Montcalm.

62

 

 

Figure 4. Processed beans of the cultivar California Dark Red Kidney 82.

63

 

 

 

 

Figure 5. Processed beans of the cultivar California Early Light Red Kidney.

After weighing, the cans were filled with boiling brine (1.25 % sodium chloride,
1.57 % sucrose, 100 mg-L‘l calcium). The filled cans were exhausted in a water-filled
exhaust box at 88 C for 5 minutes, sealed and cooked in a retort without agitation for 45
minutes at 116°C and 10.4 x 10 Pa (15 psi). After cooking, the cans were cooled under
cold running tap water and stored inverted for a minimum of 2 weeks at room
temperature.

The processed beans were placed in Styrofoam® containers for evaluation. A
team of personnel (7-12 persons) used a 7-point scale to rate the general appearance
(APP) and degee of splitting (SPLT) of each sample. The scores given by the evaluators
were averaged for each sample. The scale used for evaluation was as follows:

1 = very undesirable

2

moderately undesirable

I»
ll

slightly undesirable

4 = neither desirable nor undesirable

M
II

slightly desirable

6

moderately desirable

7 = very desirable
Identification of RAPD Markers

Samples of DNA from the parents and a subset of RILs of each population were
obtained fiom Dr. Kenneth Grafton (North Dakota State University). Seeds from the
remaining RILs were planted in the geenhouse at Michigan State University. DNA fiom
five plants of each parent and each RIL was extracted using the protocol reported by

Walters et a1. (1997).

6S

Identification of putative markers was first attempted by initially screening the
parents for polymorphisms. RAPD primers obtained fi'om Operon Technologies
(Alameda, CA) were screened against the three parents (MCM, CDRK 82 and CELRK)
to identify those that generated polymorphic bands. The primers that amplified
polymorphisms were then used to screen all the RILs in Population 1. This approach
proved to be inefficient and time-consuming. To improve the efficiency of the
procedure, selective genotyping was used for the rest of the study. For the latter
approach, five RILs from Population 1 that had the most desirable and five RILs that had
least desirable canning quality were selected. More than five lines in each DNA bulk was
considered too many and, if used, may result in the bulks not representing the extremes in
canning quality necessary for selective genotyping. Less than five lines were considered
to result in bulks that may differ not only in canning quality but in other traits as well,
which were not the interest of this study.

The choice of the five RILs for each bulk was based on the canning quality scores
of the RILs in the following environments: Mich-1996, Minn-1996, Mich-1997 and
Mich-1998 (Table 15). For each of these environments, the lines with the most and least
desirable canning quality were determined, based on their APP and SPLT scores. The
scores averaged over the four environments were also considered. The data was
compared across environments to identify lines, which consistently were the most and
least desirable in APP and SPLT scores. RILs 118-90, 118-89 and 118-97 were
consistently in the goup with the top 25% in scores for APP and SPLT in all four

environments from 1996 to 1998. In addition, canned beans from each RIL were visually

66

Table 15. Scores for appearance of processed beans of the Population 1 RILs that
were used in the DNA bulks to screen RAPD primers for polymorphism.

Appearance Rating
RIL Mich 1996 Minn 1996 Mich 1997 Mich 1998 Overall‘I

Lines with desirable canning qualin

118-90 6.0 6.1 6.3 6.2 6.1
118-89 4.7 3.5 4.7 4.9 4.4
1 18-97 4.4 4.6 4.3 4.0 4.3
1 18-60 4.2 4.7 4.0 4.4 4.3
118-73 4.5 4.6 4.0 3.9 4.2

Lines with undesirable calming qualityb

 

118-31 2.0 2.0 3.3 3.2 2.6
118-08 2.3 3.0 3.0 2.1 2.6
118-64 2.4 2.2 2.4 2.9 2.5

118-98 3.1 2.0 . 2.8 1.9 2.4
118-51 2.1 1.7 2.6 3.2 2.4
Parents

Montcalm 3.6 3.8 4.8 4.3 4.1

CDRK 82 2.0 - 2.5 2.3 2.3

Population Mean 3.2 3.0 3.7 3.6 3.4

CVc (%) 18.1 20.0 14.7 24.5 19.7

 

' Overall - averaged over the four environments

b Canning quality rating scale: 1 = very undesirable; 4 = neither desirable nor
undesirable; 7 = very desirable

b CV - Coefficient of variation

67

compared to verify that the lines chosen for the bulks represented the two extreme
phenotypes in canning quality.

DNA from the five RILs with desirable canning quality was bulked for the RAPD
analysis at a final concentration of 10 ng-ul". The same procedure was conducted with
the DNA fiom 5 lines with undesirable canning quality. Polymerase chain reaction
(PCR) amplifications of the bulked DNA were conducted in 20p] reactions containing 1X
buffer (20 mM Tris-HCl, pH 8.4; 50 mM KCl), 3 mM MgC12, 0.2 mM dNTPs, 20 ng
total genomic DNA, 20 ng primer and 1 U T aq polymerase fi'om Gibco BRL. The
RAPDS were amplified in a Perkin Elmer Cetus DNA Thermal Cycler 480 or a
Progarnmable Thermal Controller PTC-100 (MJ Research, Inc.). The progam was for 3
cycles of 1 min at 94°C, 1 min at 35°C, 2 min at 72°C; 34 cycles of 10 s at 94°C, 20 s at
40°C, 2 min at 72°C, with a third segnent extension of 1 s per cycle; and a five-min
extension at 72°C. PCR products were resolved with 100-bp and 1 Kb DNA ladders
from Gibco BRL on a 1.4% agarose gel on 1X TAE buffer.

Five hundred fifty-seven single decamer primers were screened, 107 of which
generated markers that are part of the core linkage map reported by Freyre et a1. (1998).
Markers generated from the primers were labeled with ‘O’ (Operon) to indicate the
commercial source of the primers, a letter and number indicating the kit and primer label
as used by Operon Technologies, and a number indicating the molecular size (bp) of the
marker band. The 557 primers were gouped into three sets. The first set of primers, 148
in number, was initially screened against the parents MCM and CDRK 82. After
amplification, 12 of these primers showed polymorphic bands between the two parents.

The second set, composed of 341 primers, was screened using the bulked DNA and after

68

amplification, 23 showed polymorphic bands. The third set, composed of 68 primers,
was screened simultaneously using DNA from MCM and CDRK, and from the bulked
lines; 12 of these revealed polymorphisms. A total of 47 primers revealed
polymorphisms and were used to amplify DNA from each parent and the individual
Population 1 RILs used in the bulks. Of these 47 primers, 17 were selected based on the
segegation of the bands among the 10 lines used as bulks for canning quality traits APP
and SPLT (5 with desirable and 5 with undesirable scores). Ease of scoring the bands
was also a selection criterion. Population 1 was then scored for the presence or absence
of marker bands for these 17 primers that appeared to exhibit polymorphism between the
DNA bulks selected on the basis of canning quality. The segegation ratios of these
markers in the population were determined. The markers were analyzed for linkage and
for significant associations with APP and SPLT.

Population 2 was scored for the presence or absence of the marker bands that met
the following criteria: a) segegation according to a 1:1 ratio, and b) either significant
correlation with APP or SPLT in Population 1, or linkage with markers that were
sigrificantly associated with these traits in Population 1. Eleven markers met these
criteria. Individual markers and composites of markers sigrificantly associated with APP
and SPLT were used to select lines from the second population. The canning quality
scores of these selected lines were then determined.

One of the markers identified initially, 018.1600, appeared to be identical to a
RAPD marker in linkage goup B8 of a core map constructed in the population BAT93 x
Jalo EEP558 (Freyre et al., 1998). BAT 93, Jalo EEP558 and the parents of the kidney

bean populations were amplified and resolved together in agarose gels to determine if the

69

markers mapped by F reyre et a1. (1998) and the markers detected in the kidney bean
populations were the same. Other RAPD markers in linkage goup B8 were also
analyzed for linkage to 018.1600 and for associations with APP and SPLT in the kidney
bean populations.

To determine if the markers reported by Walters et a1. (1997) to be sigrificantly
associated with canning quality in navy bean were associated with canning quality traits
in kidney bean, DNA samples from the three kidney bean parents and the navy bean
parents were amplified using the primers and amplification conditions reported by these
authors (Walters et al., 1997). The amplification products were resolved side by side on
an agarose gel to identify the markers.

Statistical Analysis and Estimation of Heritability

All data were subjected to an analysis of variance (ANOVA) appropriate to a
randomized complete block desi g1, with genotypes considered to be random, and years
and environments (year-location combinations) as fixed variables. The SAS progam
(SAS Institute, Cary, N.C, 1998) was used for the analysis. Sigrificance levels were set at
01 = 0.05. Analyses were conducted according to the following:

Analysis 1 - separate analysis for each experiment i.e., Michigan in 1996, 1997, 1998,
and 1999; Minnesota in 1996; and North Dakota in 1999.

Analysis 2 - combined data for Michigan over the years, 1996, 1997, 1998, and 1999.

Analysis 3 - combined analysis of all experiments such that years and locations were
treated as environments; only the parents and RILs of each population were

included in this analysis.

70

Box-plots of the data in Analysis 1 were constructed to provide a visual
comparison of the ranges, means and median values in the different environments. Box-
plots are interpreted as follows (Schabenberger, 1997):

a) mean - represented by (+)

b) median value - located by the line dissecting the box

c) first (Q1) and third (Q3) sample quartiles - determine the dimensions of the box. In an
ordered data set, 25% of all observations are smaller and 75% are larger than Q); 25% are
larger and 75% are smaller than the third quartile Q3. The difference between Q; and Q3
is called the inter-quartile range (IQR).

d) whiskers - represent values within 1.5 x IQR from each end of the box

e) extreme values or outliers - represented by (0) or (‘). Mild outliers (o) are
observations beyond the whiskers but less than 3 x IQR fi'om the respective end of the
box. Extreme outliers (‘) are data more than (3 x IQR) fiom each end of the box.

To estimate heritability, two replications of the data from the RILs in analyses 1
were used. Heritability was estimated on a progeny mean basis (Fehr, 1987) as follows:

H2 = 6’E = oz,
62, 0'2,er + ozylv + 0'28

 

where: 02, = genotypic variance
02, = total variance among RILs compared in r replications and v
environments (r = 2, v = 6)
02¢ experimental error
023,, = variance due to genotype x environment interactions

Confidence intervals for heritability estimates were derived according to Knapp et
a1. (1985). Correlations among the traits for each environment were determined using the
progam proc corr in SAS (SAS Institute, Cary, NC, 1998). Single-factor ANOVA was

used to detect significant associations between each marker locus and the canning quality

71

traits. Chi-square tests on the segegation ratio of the putative markers were conducted
for the two populations. Linkages between the markers that segegated according to a 1:1
ratio were determined using MAPMAKER (Whitehead Institute for Biomedical
Research, 1992). Linkage was considered sigrificant if the logarithm of odds (LOD)
score was 2 4.0.

Individual markers were analyzed against the scores of Populations 1 and 2 for
APP and SPLT in each environment, and for the APP and SPLT scores averaged over all
environments. The SAS progam proc gIm was used, with p = <0.05 for acceptance of
marker-trait associations. Markers also were gouped together, based on the results of the
linkage analysis, and analyzed for associations with APP and SPLT. The marker
. composites were as follows:

A - all the markers, which individually were sigrificantly associated with APP
and SPLT or which were linked to sigrificant markers (OY7.850, 0Q14.950, 0P15.1150,
0AGlO.l650, 0A17.4000, 018.1600, 0U20.1150, 0AH17,700, OGl7.1300,
0AN16.3000 and 0H18.1000)

B - all the markers in linkage goup M1 (OY7.850, 0Ql4.950, 0P15.1150,
0A010.1650, 0A17.4000, 018.1600 and OU20.1150)

C - all the markers in linkage goup M2 (0AH17.700, OGl7.1300, OAN 16.3000
and 0H18.1000)

D - one marker each fi'om M1 and M2 (0P15.1150 and 0G17.1300)

E - flanking markers from linkage goup Ml-l (OY7.850 and OU20.1150)

F - flanking markers fi'om linkage goup M1-2 (OY7.850 and 018.1600)

G - flanking markers from linkage goup M2 (0AH17.700 and 0H18.1000)

72

The composites of markers were used to select RILs from both populations. In
addition to these marker composites, two additional goups of markers, M1+Gl 7 and
M1+AN 16, were tested. Group M1+Gl 7 was composed of the seven markers in linkage
goup M1, and marker 0G17.1300 from linkage goup M2. Group M1+AN 16 was
composed of the M1 markers and marker OAN 16.3000. The APP and SPLT means of
the selected lines were determined. Putative markers, which were most effective as
indicators of desirable canning quality in kidney beans, were identified.

Images in this thesis are presented in color.

73

RESULTS

Three components of canning quality were evaluated and analyzed for each
population, as follows: hydration coefficient (HC), appearance (APP) and degee of
splitting (SPLT) of the canned beans. RAPD primers were screened to identify markers
associated with APP and SPLT. Insufficient data for HC was obtained for both
recombinant inbred populations planted in Minn-1996. CDRK 82 was not processed in
the 1996-Minn and 1999-NDak experiments due to insufficient seed. Eleven RAPD
markers, in two linkage goups, were identified to be Siglificantly associated with APP
and SPLT.

Evaluation of Canning Quality

Mean squares for genotypes were sigrificant for the three traits in both
populations in all analyses (Tables 16 and 17). Years and genotype x year interactions in
Analysis 2 (1996, 1997, 1998 and 1999 in Michigan) were significant for HC, APP and
SPLT for both populations (Tables 16 and 17). When the locations and years were
confounded and treated as environments (Analysis 3), environment effects and genotype
x environment interactions were sigrificant for the three traits in both p0pulations (Tables
16 and 17).

Population 1 means for APP and SPLT were similar in all environments and
ranged fiom 2.8 to 3.7 and 2.8 to 3.6 for APP and SPLT, respectively (Table 18). Means
for these traits in Population 2 were similar to Population 1 (Table 19). Coefficients of
variation (CV) for HC were very low (<2.0%) in both populations in all environments,
indicating no variation in soaking properties among the bean samples. For both APP and

SPLT, CVs were highest in Mich-1998 in both populations (Tables 18 and 19). The box

74

Table 16. Siglificance levels for main effects and interactions for canning quality traits
of Population 1 entries. Data analyses were according to individual experiments, years,
and environments (location and years confounded).

 

Source of Variation Hydration Appearanceb Degee of
Coefficient” Splittinf’L
Mean guares

 

Data Mgis number and location-year description
1 - Individual experiments

Michigan 1996: Genotype ** u in
Minnesota 1996: Genotype -' u u
Michigan 1997: Genotype "' ”I 99
Michigan 1998: Genotype ** u in
Michigan 1999: Genotype ** *9 99
North Dakota 1999: Genotype ** N u

2 - Michigan (1996, 1997, 1998, 1999)

Genotype ** *‘ *‘

Year it it 0*

Genotype x Year " " "
3 - Locations and Years Confounded, and Treated as Environments

Genotype 99 99 99

Environment “ ” "

Genotype "' Environment ” *" “

 

' - Insufficient data for HC was obtained in Minn-1996
b "" - Sigrificant at 0.05 level of sigrificance; nS - not significant

75

Table 17. Siglificance levels for main effects and interactions for canning quality traits
of Population 2 entries. Data analyses were according to individual experiments, years,
and environments (location and years confounded).

 

 

Source of Variation Hydration Appearanceb Degee of
Coefficient"b Splittingb
Mean §guares

 

Qata galfiis number and location-year description
1 - Individual experiments

Michigan 1996: Genotype ** u 44
Minnesota 1996: Genotype -‘ u in
Michigan 1997: Genotype ** u on
Michigan 1998: Genotype ** u in
Michigan 1999: Genotype ** M u
North Dakota 1999: Genotype ** an n

2 - Michigan (1996, 1997, 1998, 1999)

Genotype .. ** **
Year *0 ** **
Genotype x Year ** “ **

3 - Locations and Years Confounded, and Treated as Environments

Genotype u 99 - u
, Environment " *"‘ “
Genotype "‘ Environment " "”" "

 

' - Insufficient data for HC was obtained in 1996-MN
b """ - Siglificant at 0.05 level of Siglificance; nS - not sigrificant

76

Table 18. Hydration coefficient and scores for appearance and degee of splitting of

processed beans of Population 1 entries. Data analyses were conducted individually for

each experiment (Analysis 1).

Environment Hydration Coefficient Appearance Degrpe of Splitting
Mean CV (%)”I Mean CV (%)'I Mean CV (%)‘I

Mich 1996 2.21 1.98 3.2 18.1 2.8 22.1
Minn 1996 -° -° 3.0 20.0 2.6 21.6
Mich 1997 2.10 1.57 3.7 14.7 3.6 16.0
Mich 1998 2.27 1.03 3.6 24.5 3.5 25.7
Mich 1999 2.15 1.75 3.7 17.4 3.6 17.2
NDak 1999 2.20 1.18 2.8 18.5 2.9 18.9

 

‘ CV - coefficient of variation
D - Insufficient data was obtained

Table 19. Hydration coefficient and scores for appearance and degee of splitting of

processed beans of Population 2 entries. Data analyses were conducted individually for

each experiment (Analjsis Q.

Environment Hydration Coefficient Appearance Deggee of Splitting
Mean CV (%)a Mean CV (%)' Mean CV (%)'I

Mich 1996 2.24 1.61 3.4 18.4 2.8 21.5
Minn 1996 -° -° 3.1 17.3 2.8 20.9
Mich 1997 2.19 1.37 3.7 13.7 3.7 13.2
Mich 1998 2.26 1.43 3.2 22.3 3.2 21.2
Mich 1999 2.24 1.76 3.4 17.7 3.4 18.4
NDak 1999 2.17 1.15 3.2 16.6 3.2 15.9

 

' CV - coefficient of variation
° - Insufficient data

77

plots in Figures 6 and 7 Show several outliers for APP and SPLT scores in the two
populations. These outliers are the RILs with high scores for canning quality traits.

When the Mich data for Population 1 were combined over the four years of the
study (Analysis 2) (Table 20), the scores for both APP and SPLT were Siglificantly lower
in 1996 than the other three years. The box plots of data from Population 1 for 1996,
1997, 1998 and 1999 in Michigan (Figure 6) illustrates these results. There were no
Significant differences between the years 1997, 1998 and 1999 for APP and SPLT (Table
20). The APP and SPLT scores for 1996 were 3.1 and 2.7, respectively. For 1997, 1998
and 1999, the APP scores were similar (3.6 to 3.7) as were the SPLT scores (3.5 to 3.6).
111 Population 2 (Table 21), the entries planted in 1997 had sigrificantly higher mean
scores for APP and SPLT than those planted in 1996, 1998 and 1999 in Michigan
(Analysis 2). For both APP and SPLT, the mean score for 1997 was 3.7. The box plots
for the data from Population 2 for these four years in Michigan (Figure 7) illustrate these
results.

There were Siglificant differences between some environments for Populations 1
and 2 for HC, APP and SPLT (Tables 22 and 23). In Population 1, HC had the highest
values in the Minn-1996 and Mich-1998 environments (Table 22). Both APP and SPLT
scores, which ranged from 3.5 to 3.6, were highest in Mich-1997, Mich-1998 and Mich-
1999, with no sigrificant differences among these three environments. In Population 2
(Table 23), the highest scores for APP and SPLT were fi'om the Mich-1997 environment
(3.7 for both APP and SPLT), which was sigrificantly different from the other five

environments.

78

 

 

0 - extreme outliers

l- mild outliers

 

 

" -median

 

 

A. Appearance (APP)

am a 90'.
o I u
e w IIIII". "IIIII
g 000 .u .
. . I .
+ ‘ 90'.
O It'll 9
. O o
. I .
III III c Q 0 III III
- O o
. o .
O IOII O
0110' O
. o .
. c .
0° 0 OII IIII . O I . III IIIO
. o .
. I .
OIOI .9
O IOII O
. u .
. I .
00° IIIII. O .IIII
. o .
a D n
9 I‘ll O
OI 0.116
. I .
. I .
o a ll'l II. . ollllllll
. I .
. I .
OI fill.
." .l.
I .
I o
o OIIIIIII O . .IIIIIIIII
I o
O o
0" CI.
9 O O O O O
7 6 5 0 3 2

 

O

. Environment

 

 

 

mm .
» oIIIII" u. "IIIII
”m u u u
. OI. 9
mm 9IIoI9
. . 9
7m IIIIIII .m mIIIIII
"M .II..Ia
m8 OIOIO
aw " n u
u .w o 0 o 0" IIII " fl " III III 0
"M ”1.1”
m7 III.
a w IIII " " "III III
Tm once I. u
"M XII»
m6
“ “1.14
um i. .. m
mM T ”1.1”
u L
"6 P
mm «(Si “111.1.
«.m g 00 IIIIIIIIm 9m "IIII
WM .m ”III..I.
" h
. P
S
cm '9 O O O O 9 O
m 7 6 5 G 3 2 1
c
D
B.

Environment

0-....-0.0-.--.-OCOOOQOO.-.-0.0--.D-0..-.O.....I...-..-.-.OC..

.O----.--..-..--...

Mich98 Mich99 NDak99

Minn96 Mich97

Mich96

Figure 6. Data Analysis 1 - Box plots of scores for a) appearance and b) degee of
splitting of processed beans of Population 1 RILs, parents and checks, planted in

Michigan, Minnesota, and North Dakota from 1996 to 1999.

79

 

" - median
Environment

 

 

0.

0 — extreme outliers
l- mild outliers

+-mean

.

Legend:

 

 

.U--.

 

Mich99 NDak99

.O--...-......-..O-C.-..-.--.

'9

.IIIII.
.O.-.
I
I
I
|
|
CO...-

I
.O.-....-..-

Minn96 Mich97 Mich98

Mich96

 

B. Degree of Splitting (SPLT)

A. Appearance (APP)

OIIII
I

II IIIIIIII ll . O
I
.

Ollll

O llll

.IIOII.

O llll

.
O

O
I
ovoe IIIIII IIIIII.

.
9

Environment

OO...--O.-..-.....--.--..--

Mich99 NDak99

80

Minn96 Mich97 Mich98

Mich96

.O...................-.-.....-....

9
O

 

O
9
9

planted in Michigan, Minnesota, and North Dakota from 1996 to 1999.

Figure 7. Data Analysis 1 - Box plots of scores for a) appearance and b) degree of
splitting of processed beans of Population 2 RILs, parents and checks,

Table 20. Hydration coefficient and scores for appearance and degree of splitting of
processed beans of Population 1 entries. Data analyses were conducted to compare
individual years in Michigan iAnalysis 2).

 

 

Year Hydration Appearancea Degree of
Coefficienta Splitting'I
1996 2.21 b 3 . l b 2.7 b
1997 2. 10 d 3.7 a 3.6 a
1998 2.27 a 3 .6 a 3.5 a
1999 2.15 c 3.7 a 3 .6 a
Mean combined over years 2.18 3.5 3.4
Coefficient of variation (%) 1.62 18.9 20.3

 

' - Means with the same letter are not significantly different by Fisher's LSD (0.05).

Table 21 . Hydration coefficient and scores for appearance and degree of splitting of
processed beans of Population 2 entries. Data analyses were conducted to compare
individual years in Michigan (Analysis 2).

 

 

Year Hydration Appearancell Degree of
Coefficient' Splitting'
1996 2.24 b 3.4 b 2.8 d
1997 2.19 c 3.7 a 3.7 a
1998 2.26 a 3.2 c 3.2 c
1999 2.24 b 3.4 b 3.4 b -
Mean combined over years 2.23 3.4 3.3
Coefficient of variation (%) 1.57 18.2 18.4

 

' - Means with the same letter are not significantly different by Fisher’s LSD (0.05).

81

Table 22. Hydration coefficient and scores for appearance and degree of splitting of
processed beans of Population 1 parents and RILs grown in Michigan, Minnesota and
North Dakota from 1996 to 1999. Analyses were conducted to compare year-location
combinations, treated as environments (Analysis 3).

 

 

Environment Hydration Appearancea Degree of
Coefficient‘I Splitting‘I
Mich (1996) 2.20 b 3.1 b 2.7 bc
Minn (1996) 2.27 a 3.0 b 2.6 c
Mich (1997) 2.10 e 3.6 a 3.5 a
Mich (1998) 2.27 a 3.6 a 3.5 a
Mich (1999) 2.15 d 3.6 a 3.6 a
NDak (1999) 2.19 c 2.7 c 2.8 b
Mean combined over
environments 2.19 3.3 3.1

Coefficient of variation (%) 1.50 19.2 20.6

 

' - Means with the same letter are not significantly different by Fisher's LSD (0.05)

Table 23. Hydration coefficient and scores for appearance and degree of splitting of
processed beans of Population 2 parents and RILs grown in Michigan, Minnesota and
North Dakota fi'om 1996 to 1999. Analyses were conducted to compare year-location
combinations, treated as environments (Analysis 3).

 

 

Environment Hydration Appearance'I Degree of
Coefficient' Splitting'l
Mich (1996) 2.24 c 3.4 b 2.8 d
Minn (1996) 2.28 a 3.2 c 2.8 d
Mich (1997) 2.19 d 3.7 a 3.7 a
Mich (1998) 2.26 b 3.2 c 3.2 c
Mich (1999) 2.24 c 3.5 b 3.4 b
NDak (1999) 2.17 e 3.1 c 3.2 c
Mean combined over
environments 2.22 3.3 3.2
Coefficient of variation (%) 1.51 17.9 18.5

 

' - Means with the same letter are not significantly different by Fisher's LSD (0.05)

82

RILs with high APP and SPLT scores in Populations l and 2. CDRK 82, one of the
parents of Population 1, was not processed and canned in two environments, Minn-1996
and NDak-1999, due to insufficient seed. In the other four environments, CDRK 82,
planted with Population 1, had undesirable APP scores (2.0 to 2.8) (Table 24). The SPLT
scores of CDRK in these four enviromnents ranged fiom 1.5 to 2.6. MCM had APP
scores of 2.6 to 4.8 and SPLT scores of 3.4 to 4.9. In the four environments where both
parents were evaluated, large differences were not observed between the mean scores of
the RILs and the mean scores of the parents, either for APP or SPLT (Table 24). In each
environment, the RILs with the ten highest scores for APP in Population 1 were identified
(Appendix Tables A.5 and A.6). The mean APP and SPLT scores of these 10 RILs were
higher than the mean scores of the parents in all of the environments (Table 24). Except
for Mich-1997, the mean APP and scores of the 10 RILs were also higher than the mean
scores of MCM, the parent with the more desirable canning quality. Furthermore, the
mean APP and SPLT scores of these 10 RILs were higher than or comparable to the mean
scores of the check varieties in all the environments.

In Population 1, one RIL, 118-90 (Figure 8), consistently had the highest score
among the RILs for both APP and SPLT in all six environments (Table 24; Appendix
Tables A5 and A.6). The scores for RIL 118-90 had ranged fi'om slightly to moderately
desirable (APP = 4.6 to 6.3; SPLT = 4.6 to 6.5). The canned beans fi'om this RIL were
generally intact, had few splits, and the color of the cooked beans was judged acceptable
for the market class. For comparison, RIL 118-51, a line with undesirable canning quality
is shown in Figure 9. RIL 118-51 showed numerous split beans, sloughed seed coats and

pieces of cotyledon in the brine. Except for Mich-1997 and NDak-1999,

83

.. . @3568

 

 

 

 

 

 

 

 

 

o.m 9m o.m 3“ m6 m6 0% Id «mm Id Wm WM 83352
a- .. o.m Id Wm QN Wm m.~ .. ... m._ c.~ .Nw vans
_ couflnmom mo napalm
s- c- owe. Ii c- c- c- a. mum e6 Nflv v6 3-x:
a- a- N. v . I v p- c- a- a- m m n v o m N v 3-x:
o v Id 2. c- c- a- :u ma. ~.m ad a. c- 3%:
c- c- c- c- _..n aé wd Ev c- c- is 5v owIw:
v m Wm .né 5v 5v e... c- a- c- c- c- 3-x:
a- «av. vi c- c- 0% Ev c- a- 3d e6 3-2 _
3“. m6 c- c- c- c- Nd ed. a. c- 3. Id 3in
3358:8365
5m 2: .«o 8:: 3 8.80m Him 93 an? “monar— o_ 2: 5; 98cm 05 9 :oEEou 39$ 2M5 32
m6 v.9 w.m Em m6 N6 m6 n6 «.0 fig me :6 CQIwZ
35:80:50 5m an E 888 Him 93 nE< “we: .3 o_ of £3; :8 2: 8 cog—E8 3.5 35 dd
.Smm mm< Hiram mm< Hiram a: Ham mm< Hun—m a: Ham mm< commmuuo<
cmoeoum “a: Emma mag
32 mg— wmfl 32 eg— cog
can—Z 5:2 cog no: :32 :32
Hangobém

 

.385 2: Ea .3532 2: .32 mew—8m SHINE 2 05 .32 2: =a
am .antoaxo 2: «o 888 82a of 5? $8.55on 333:5 5m 28 85 a 32 mo 3on mecca “mos—ma R: 2: 5
Banana :2: 32 98 3:28 _ cog—anon go 289 338830 maxim mo uocmou 93 85325.“ no.“ «Boom .3 £93.

84

A: :0.“ 05 0%. 20: E 00:73 c
00350 0000 05:05:05

2 0:: 3:059:50 0005 E .0050 00: 53 mm van—U 0050 000.55.. 05 qum :5 an? 3.: 00.500 508 oz 0

3 .3 00:00:05 05 00.5.: .0: 0005 3:058:60 :000 :2 .0253 0.53 8.00:0 05 :0 :02 9
.:30:0 05

00.500 32$: 2 05 5:5 00:: he 95:» 05 w:o5 0.53 add 05 80:3 8:08:835 05 5 00.800 Ham :5 mm< 3:0 9
00—950 .0000 80658::

8 2:0 mam—-592 :5 003-522 E 0050 :5 0000008: 00: - 3m >055— :3: 05G 050.508 mm van—U .

..

o.w_ mi N: v.5 fimm méu Q:— h.: 0.5 adu _.NN fia— Ac\cv >0
_._ —.— «2 MA I: IA ~.— ~.— N: M; N: ~.— 28.8 Om..—
h.m sum 0.: a... a4. Qm 0.: «a. o6 ma. Nd m4. mind “mom
IN EN 9m ed Wm 9n Wm ed 0N fin Nam —.n a =<

a- m.m v.n Wm WM Nun sum 0. ON a." 3:05.:
N6 ed. ad hi V4. vi Ev «a. 0.. mé v... 00:05.; 0.00:0

ad ad 04m fin Wm 9m 9m Em 9N :6. IN fin 50.: 0:08:0me
>0 :i5 Gm: 3.: 0023 v5 0152
05 5 add 2. :0 8:05 9.00:0 :08. x0 05 .«o 0502

 

 

 

 
 

M5500 amonwm: 2 05 Ed :33:

 

.Smm mm< Ham a: .Smm .54 .Smm a: .Smm mm< H55 9:. :0_00000<
c0208 :0: 3:0:aiwi:_:5o

 

 

33 33 III: 32 :3— ca:
0:52 :022 :22 :32 5:2 :022
.:0E:o:>~fl

 

.85528 .00 23.

85

 

 

 

 

 

Figure 8. Processed beans of recombinant inbred line 118-90,
from a cross between Montcalm and California Dark Red Kidney 82.

86

 

 

Fat»;

fir ,,. .

 

 

Hoff

 

 

Figure 9. Processed beans of Population 1 recombinant inbred line 118-51,
from a cross between Montcalm and California Dark Red Kidney 82.

87

RIL 118-9O had a higher score for both APP and SPLT than the parents and checks
(Appendix Tables A.5 and A.6).

In addition to RIL 118-90, several RILs were among the 10 RILs with the highest
APP scores in more than one environment. Seven RILs had high APP scores in three
environments, with scores ranging from 3.5 to 4.9 (Table 24). Of these seven, two RILs
118-09 and 118-89, were among the 10 highest scoring RILs for APP and SPLT in
Michigan over three years. RIL 118-O9 had APP scores ranging from 4.4 to 4.7 and
SPLT scores ranging from 4.4 to 4.6. RIL 118-89 had APP scores ranging from 4.7 to 4.9
and SPLT scores ranging from 4.4 to 5.1. APP ratings for the five other RILs, 118-05
(3.9 - 4.6), 118-66 (3.5 - 4.7), 118-93 (3.8 - 4.5), 118-95 (4.2 - 4.8), and 118-97 (4.4 - 4.8)
were among the 10 highest scoring RILs in three different year-location combinations. In
most of the environments, the lines with the 10 highest APP scores in more than two
environments were comparable to or better than MCM (Table 24). The APP and SPLT
scores of all the RILs and the parents of Population 1, and the checks in each environment
and averaged across all six environments are shown in Appendix Tables A.5 and A.6. In
each environment, one or more RILs had higher scores for both APP and SPLT than the
best check variety.

In Population 2, CELRK had higher APP and SPLT scores than MCM, in Mich-
1997 and in NDak-1999 (Table 25). MCM is generally known to have a more desirable
canning quality than CELRK. In all environments, the parents had higher mean scores
for APP and SPLT than the RILs. These differences were significant only for Michigan
in 1997. The mean APP and SPLT scores of the 10 RILs with the highest scores for APP

and SPLT were higher than the mid-parent scores in all environments except in

88

. . .:0:E::00

 

 

0:0 0:: .:0 000.0%: 00:000 Ham :5 05.0. :00: E o— 0:: :55 :0: 0:: 0: 008800 05.5 :0:: 0:2

 

B. VE:

$-0— :
0358:0035

 

 

 

 

0:. 0... .- .- Z. 3. 0.0 0.0 .- .- 0.0 0.0 02 20: 00-0: _

0.0 0... 0.0 0.0 .- .- .- .- 3. 0... 3. S. 08 20: 00-0:

0.0 0.0 3 0... .- .- 3. 0... 0.0 0... .- .- 00.. 0:00 00-0:

.- .- S. 0.0 .- .- 0.0 0.0 0:. 0... 0.0 a... B. 20: 00-2:

.- .- 3. 3. 3 3. is 0... o- .- 0.0 0... B: :0: 3-0:
o35:80:35

00.0 0:: .:0 :00: E 00:000 Him :5 05¢. :00: :: o: 0:: :5: 00: 0:: 0: 008800 05.5 :3: 0:2

2. 0... 0.0 . 0.0 .- .- 0... :0 0.0 0.0 N0 0.0 B: v:00 V0-0:
U350.800.0000

 

0:0 0:: .:0 0.»: :: 00:000. .Smm :5 A540: :00: E o— 0:: 5.5» :0: 0:: 0: 02:88 05.5 :0:: 0:2

 

 

.Smm 03¢. Ham 0::. Han—m 07—4.. Han—m 0::. him a: Ham a:
000:000 :00: 5:03 w50

 

 

000: 000: 000: $0: 000: 000:
V:02 5:2 6:2 5:2 8:2 00:2
:l5:::0:>fi_

0200000.:~

 

0:55 0:: :5 0:55: 0:: .042 M5800 :00:m::
o— 0:: 0.5: :0 508:5on 0:: .:0 000000 50:: 0:: :5: 0:55:00 50:36:: 0.»: :5 80.: 000:: :: 0:2 .:0 98E m::.:000
50:3: 5: 0:: 0: 550050 :0:: 32 N 5:03:00: .:0 50: 500000:— .:0 39:00 .:0 00:w0: :5 005000050 00.: 00:00m .3. 050:.

89

.002 00.: 000 20.0: 00.0: 0:: 000:3 0

00:00:00 :000 6000:0000: 0: 0:: 000: 00: 0:00:09 5002 :0 0:05:00 :00 0000000000 :0: 0003 0.0000 n u
030:0 000 0.0000 0000000 00:00: :0: 0:: 0:00:00 0003 000000 0.00”: 0:: 000:3 0000000000300 000:: 0: 000000 0:: 00:0 - o

08:5: 80 200.0 0:00 05.00000 - 02.000 ..
:00 ::m: :5 0:00: .00 00:00:00 0 ”00:00 :000 30000580000070 .

 

 

 

 

 

 

0.0 0 :60 :.w: 0.0.0 ~.:N n. «N ~.n0 0.n 0 0.o~ n. 00 n.0N :.m0 0x0: >0
0.0 0.0 N: «.0 :.: :.0 o: 0.0 fl: 0.0 N: «.0 $0.0: Gm:
0.: 0.: w: 0.n N: N: 0.: 0.: n.: 0.: N: 0.: 0.00m $000000 :00:w::.0
Nn ~.n :.n :.n «.0 an 0.n 0.n 0N 0.n wd :.n 00.00”: =<
0.: 0.: 0.n 0.: 0.n 0.n 0.: 0.n 0.n N.n 0.n 0.n 0:000:00
:.m :.n 0N 0.n 0.n 0.n 0.: 0.: n.~ 0.n 0.n 0.n 000:0:00> 008:0
an ~.n :.n :.n N.n ~.n 0.n 0.n «N 0.n w.~ :.n 0000000 2000000009000

Rug Gm: 00.: 00::0M3 :00 0.00m
N.n ~.n N: :.: 0.: 0.: 0m: 0.: Nn 0.n :.n 0.: :00 000000 00:00:52
0.: 0.: Nn 0.n 0N 0.n 0.n ~.m Nu 9N 0N ~.n :00 ::w:.0 aMmAMU
N Bum—00000 .00 0:00.09:
0.40m 0.0:. 0.4.5 0000:. 0.1005 .0: Exam .0: 0.00m 0.0:: 0.40m .0.0<
0000000 :08: 0::—0:0 wfifiao
0000 000: 0000 0000 :00: :00: 00:00000<
000002 :0002 :00: :22 50:0: 000:0):
:0000000000:>:m

 

08.30080 .00 2000

90

Mich-1997. Except for the Minn-1996 environment, the differences were non-
significant. Furthermore, the 10 RILs with the highest APP and SPLT scores had
significantly higher mean scores than the check varieties in all environments except for
Mich-1997 and Mich-1998.

Seven RILs in Population 2 were among the 10 lines with the highest APP scores
in more than two environments (Table 25). Three of these RILs had the seed color and
seed appearance of dark red kidney beans while four RILs belonged to the light red
kidney bean market class. Of these seven RILs, RIL 119-34 (3 dark red kidney bean) was
in the group with the 10 highest APP scores in five of the six environments, with scores
ranging from 4.5 to 5.9. Five RILs were among the lines with high APP scores in four
environments. Among these lines, RIL 119-20 (a light red kidney line) was among the 10
RILs with the most desirable canning quality in Michigan in all four years of testing, but
was not among the 10 highest scoring RILs for APP in Minnesota or North Dakota. RIL
119-33 was in the group of 10 RILs with high APP scores in Michigan in 1996, 1997 and
1999, and in Minnesota in 1996; RIL 119-78 was in the group in Michigan from 1996 to
1998, and in North Dakota in 1999. RIL 119-69 was among the 10 highest scoring RILs
for APP scores in 1997 and 1999 in Michigan, in 1996 in Minnesota and in North Dakota
in 1999. RIL 119-72 also was common to the group in four environments: Mich-1996,
Mich-1999, Minn-1996 and NDak-1999. For comparison, MCM had APP scores beans
ranging from 3.2 to 4.9 (Table 25). The APP and SPLT scores of these and the rest of the
RILs, along with the parents and check varieties, ranked from highest to lowest APP

score, are shown in Appendix Tables A.7 and A.8.

91

Heritability estimates and coefficients of correlation between tra1_'t§. The ANOVA tables
fi'orn which the variance components estimates were calculated from mean squares for
APP and SPLT are presented in Appendix Tables A.13 to A.16. The narrow-sense
heritability estimates for APP and SPLT were similar in both populations and were
approaching high value (~0.9) (Table 26). For Population 1, the heritability estimates
were 0.83 and 0.84 for APP and SPLT, respectively. For Population 2, APP and SPLT

had a heritability of 0.85 (Table 26).

Table 26. Heritability estimates for appearance and degree of splitting of
processed beans in Populations l and 2, using data from all environments combined.

 

Appearance (CI‘) Degree of splittinL(CI°)
Population 1 0.83 (0.75 - 0.87) 0.84 (0.76 - 0.87)
Population 2 0.85 (0.77 - 0.89) 0.85 (0.78 - 0.89)

 

' - 2 replications in 2 locations and 2 years.
b - 2 replications in 6 environments
° CI — 95% confidence interval

Pair-wise correlations among the canning quality traits - HC, APP and SPLT - and
also between the canning quality traits, and yield and seed weight, are shown in Tables 27
and 28, for Populations 1 and 2, respectively. In both populations, APP and SPLT were
highly correlated in each environment. Coefficients of correlation between these two
traits ranged from 0.91 to 0.97. BC was positively correlated with APP in four
environments in Population 1 and in five environments in Population 2; coefficients of

correlation ranged from 0.24 to 0.68. HC was also correlated with SPLT in three years in

92

Michigan, in 1996, 1997 and 1999 in Population 1 and in five environments in Population
2; coefficients of correlation ranged from 0.34 to 0.63 (Tables 27 and 28).

In Population 1, both APP and SPLT were negatively correlated with yield in both
locations in 1996, and in North Dakota in 1999 (Table 27). Coefficients of correlation
between APP and yield ranged fi'om —0.43 to —0.24, and for SPLT and yield, the
coefficients ranged from —0.47 to -0.26. In Population 2, APP and yield were negatively
correlated in all six environments; coefficients of correlation ranged from —0.15 to —0.50
(Table 28). SPLT and yield were negatively correlated in five environments; coefficients
of correlation ranged from -0.18 to —0.55.

In Population 1, negative correlations between APP and seed weight, and between
SPLT and seed weight were detected in 1996 and in 1999. Coefficients of correlation
between APP and seed weight ranged from -0.37 to -0.29. For SPLT and seed weight, the
coefficients ranged from -0.42 to -0.28 (Table 27). In Population 2, APP and SPLT were
negatively correlated with seed weight in five environments; coefficients of correlation
ranged fiom —0.20 to —0.37 (Table 28). Other correlations detected were between HC and

seed size, and between HC and yield (Tables 27 and 28).

Identification of Putative Markers for Canning Quality Traits

The primers (0C5, OM10, 0N17, 0011, 004, 0P5, 0Y5, 0M11, OX3, 0016,
0Y13, 0F 5, 0N18, AND 0AC2) that generated RAPD markers associated with canning
quality traits in navy bean populations (Walters, 1995; Walters et al., 1997) were screened
against the three kidney bean parents in the current study. For four primers (0C5, 0011,

004 and 0AC2), the marker bands were not identified due to faint

93

Table 27. Significant correlations between canning quality traits, seed weight and yield
in Population I, planted in Michigan, Minnesota and North Dakota from 1996 to 1999.
Environmenta

Traitl Trait 2 Mich Minn” Mich Mich Mich NDak Coefficient of
1996 1996 1997 1998 1999 1999 Correlation

yield seed weight * "‘ * * 0.42 to 0.66
HC seed weight - "' -0.38

HC yield - "‘ * -0.39 to -0.62
APP yield “ * * -0.24 to -O.43
APP seed weight “ * -0.29 to -0.37
APP HC "' "‘ "' 0.37 to 0.68
SPLT yield * "‘ -0.26 to -0.47
SPLT seed weight "‘ * -0.28 to -0.42
SPLT HC "' - 0.36 to 0.59
APP SPLT "' "' * "‘ * * 0.94 to 0.97

 

' "' - Significant at level of significance = 0.05
b - Insufficient data was obtained for BC in 1996-MN

Table 28. Significant correlations between canning quality traits, seed weight and yield
in Population 2, planted in Michigan, Minnesota and North Dakota from 1996 to 1999.
Environment"1
Trait 1 Trait 2 Mich Minnal Mich Mich Mich NDak Coefficient of
1996 1996 1997 1998 1999 1999 Correlation

HC seed weight * - * r * -o.32 to 0.25
HC yield a - * * -o.21 to -0.58
APP seed weight * * * * * -o.2o to -033
APP yield * * * * * -o.15 to -o.50
APP HC * - r * * * 0.24 to 0.60
SPLT seed weight * * * * * * .022 to .037
SPLT yield * * * * * -0.18 to -o.55
SPLT HC * - * * * 0.34 to 0.63
APP SPLT * * * * * 0.91 to 0.97

 

' "' - Significant at level of significance = 0.05
b - Insufficient data was obtained for HC in 1996-MN

amplification products. Twelve primers - OM10, ON17, 0P5, 0Y4, OM11, OX3, 0016,
0Y13, OFS, and ON18 - did not amplify polymorphic bands in the kidney bean parents,
MCM, CDRK 82 and CELRK. Primer 0P5 detected a polymorphism among the kidney
bean parents, but the band was faint and difficult to score in the population. No further
amplifications with this primer were conducted on either Population 1 or 2.

Selective genotyping (Miklas et al., 1996) was effective in identifying markers
associated with canning quality traits in the two kidney bean populations. Of the 8
markers identified initially, 4 (1.2% of 341 primers) were screened using the bulked DNA
procedure. One marker (0.7% of 148 primers) was screened between the parents only.
Three (4.4% of 68 primers) were screened using the parents and the bulks simultaneously.
After the initial screening of the primers, the marker genotypes of the individual lines,
which composed the bulks, were evaluated to determine the segregation of the marker
bands among these 10 lines. This approach was less time-consuming and more efficient
than if the entire p0pulation were scored immediately.

Thirteen RAPD bands - 0Y7.850, 0014.950, 0P15.1150, OAGlO. 1650,
0Al7.4000, 018.1600, OU20.1150, OGl7.1300, OAN 16.3000, OH18.1000, OA7.2100
and 001700 - segregated in Population 1 according to a 1:1 ratio (Figure 10). Only these
13 bands were used for scoring the populations since the use of markers with distorted
segregation ratios increases the possibility of detecting false positive polymorphism
(Wang and Peterson, 1994). Eleven of these marker bands formed two linkage groups,
designated Ml-l and M2-1, in Population 1. Seven markers - 0014.950, 0P15.1150,
OAGlO.1650, OY7.850, 018.1600, 0U20.1150 and 0A17.4000 - comprised linkage

group Ml-l, with a total map distance of25.9 cM (Figure 11). Markers 0P15.1150 and

95

Ill-d fiuuaflu.

Cyan“ ~
.,,

 

Figure 10. Amplification of primer 0017, showing marker OGl 7.1300,
using DNA from parents and some RILs of Populations l and 2:
Primer CG] 7: Lanes 1 and 30 — 100 bp ladder; 2 — MCM;

3 — CDRK 82; 4 to 9 — some RILs of Populations l.

96

Ml-l M2-l

OY7.850 OAH17.700
9.6

9.8
0014.950
0.7
OP15.1150 / OAGl0.1650 OGl7.1300
2.1 3.8
OA17.4000

OAN16. 3000
3.6
018.1600

12.5
9.8
OU20.1150

OH18.1000

Figure 11. Linkage groups detected in Population 1 (MCM x CDRK 82).
Ml-l (OY7.850, 0014.950, OP15.1150, OAGlO.1650, OA17.4000,
018.1600, OU20.1150), total map distance = 25.9 cM;
M2-1 (CAI-117.700, OGl7.1300, OAN 16.3000, 0H18.1000), total map distance = 26.1
cM.

97

OAGlO.1650 had a map distance of 0 cM between them, indicating no recombination
between these two markers in this population. Four markers - OAH17.700, 617.1300,
OAN 16.3000 and OH18.1000 - comprised linkage group M2-1, with a total map distance
of26.l cM (Figure 11).

In Population 2, two linkage groups, composed of the same markers as the linkage
groups in Population 1, were detected (Figure 12), designated Ml-2 and M2-2. The same
seven markers in Ml-l (Population 1) comprised M1-2, but were in a different order in
Population 2. This map had a total map distance of only 6.5 cM. Markers U20.1150,
OP15.1150 and 0014.950 were very closely linked (map distance = 0.0 cM), and had a
distance of 0.7 cM to OAGlO.1650. Linkage group M2-2 (total map distance = 27.4 cM)
had the same four markers in the same order as in M2-1 (Population 1).

In order to identify a possible location of the M1-1/M1-2 maps relative to the core
map of the bean genome (Freyre et al., 1998), DNA samples fiom MCM, CDRK 82,
CELRK, BAT 93 and J alo EEP558 were amplified using the RAPD primers 018 and
U20, and resolved side by side on agarose gels (Figure 13). The results indicated that the
marker 18.1500 in linkage group B8 reported by Freyre et a1. (1998) might be a length
polymorphism between these two lines, with BAT having a band 1500 bp long and Jalo
with a slightly longer band, about 1600 bp. The band that was polymorphic among
MCM, CDRK 82 and CELRK was also about 1600 bp long, indicating that marker
18.1500 reported by Freyre et al. (1998) and marker 018.1600 found in the two kidney
bean populations may be at the same locus. Marker U20.1150 was clearly the same band
that was polymorphic between BAT 93 and J alo EEP558, and among the kidney bean

parents, MCM, CDRK 82 and CELRK (Figure 13).

98

M2-2

OAH17.700
9.1
OGl7.1300
Ml-Z

OY7.850
l 1.0

2.2

OU20.1150 / 0014.950 I OP15.1150

0.7

OAGlO.1650

1.4 OAN16. 3000

OA17.4000

0.7

018.1600
7.3
OH18.1000

Figure 12. Linkage groups detected in Population 2 (MCM x CELRK).
Ml-2 (OY7.850, 0014.950, OP15.1150, OAGlO.1650, OA17.4000,
018.1600, OU20.1150), total map distance = 6.5 cM;
M2-2 (OAH17.700, OGl7.1300, OAN 16.3000, OH18.1000), total map distance = 27.4
cM.

99

8933"

.¢o.m:‘:~

 

Figure 13. Amplification products of RAPD primers 018 and OU20, showing
Markers 018.1600 and OU20.1150, respectively.

a. 018: Lanesl and 7 —1 Kb ladder; 2 — CDRK 82; 3 — MCM;
4 — CELRK; 5 - BAT93; 6 - Jalo EEP558.

b. OU20: Lane 1 - 100 bp ladder; 2 - CDRK 82;
3 — MCM; 4 — CELRK; 5 — BAT93; 6 - Jalo EEP558.

100

Identification of putative RAPD markers in Population 1. Of the 13 markers that

segregated at a 1:1 ratio in Population 1, nine were significantly correlated with APP and
SPLT scores in at least one environment (Table 29). Among these nine markers
significantly associated with canning quality traits were all the seven markers in linkage
group Ml-l (Figure 11). Two markers in linkage group M2-1, OGl7.1300 and

OAN 16.3000, were significantly associated with APP and SPLT in at least one
environment. The other two markers in linkage group M2-1, 0AH17.700 and
OH18.1000, were not significantly associated with APP or SPLT in Population 1 in any
environment.

The seven markers in linkage group Ml-l had very similar patterns in terms of the
environments in which the markers were significantly associated with APP and SPLT
(Table 29). Six of the seven markers in this linkage group were significantly associated
with APP in 1996, 1997 and 1999 in Michigan, but not in Minn-1996, Mich-1998 or
NDak-1999. Marker OU20.1150 was significantly associated with APP only in Mich-
1996. For SPLT, six markers were associated with the trait in 1996, 1997 and 1999 in
Michigan. Marker OY7.850 was further associated with SPLT in Mich-1998. Again,
marker OU20.1150 was significantly associated with SPLT only in Mich-1996. All the
markers were significantly associated with APP and SPLT scores averaged over all
environments.

No pattern was observed in the M2-l markers, OGl7.1300 and OAN16.3000.
Marker 001 7.1300 was significantly associated with APP in Minn-1996, Mich-1999 and
NDak-1999, and with SPLT in Minn-1996 and NDak-1999. Marker OAN 16.3000 was

significantly associated with APP in Mich-1999 and with SPLT in NDak-1999. Both

10]

Table 29. Coefficients of determination (R‘) of RAPD markers associated with scores
for appearance and degree of splitting of 75 RILs of Population I, planted in Michigan,
Minnesota, and North Dakota, fiom 1996 to 1999.

 

 

2W
Markers' Mich Minn Mich Mich Mich NDak All Env.b
1996 1996 1997 1998 1999 1999
Apmarance (APP)

Linkage gmpp M1

OY7.850 0.062 - 0.046 - 0.083 - 0.082
0014.950 0.073 - 0.079 - 0.065 - 0.068
0P15.1150 0.079 - 0.067 - 0.060 - 0.069
OAGlO.1650 0.078 - 0.064 - 0.062 - 0.069
OA17.4000 0.051 - 0.031 - 0.044 - 0.039
018.1600 0.072 - 0.047 - 0.044 - 0.052
OU20.1 150 0.043 - - - - - 0.038
Linkage glpup M2

OAH17.700 - - - - - - -
001 7.1300 - 0.036 - - 0.031 0.047 0.036
OAN 16.3000 - - - - 0.025 - 0.026
OH18.1000 - - - - - - -

Dggree of splitting (SPLI)

Linkage gmup M1

OY7.850 0.062 - 0.037 0.029 0.074 - 0.071
0014.950 0.092 - 0.083 - 0.079 - 0.084
OP15.1 150 0.099 - 0.066 - 0.072 - 0.083
OAGlO.1650 0.099 - 0.064 - 0.073 - 0.084
OA17.4000 0.076 - 0.026 - 0.044 - 0.045
018.1600 0.091 - 0.044 - 0.046 - 0.058
OU20.1 150 0.066 - - - - - 0.044
Linkage gmpp M2

OAI-I17.700 - - - - - - -
OGl7.1300 - 0.056 - - - 0.059 0.035
0AN16.3000 - - - - - 0.044 0.027
OH18.1000 - - - - - - -

 

' (-) - not significant at level of significance = 0.05

b All Env. - APP and SPLT scores averaged over all environments

102

OGl7.1300 and OAN 16.3000 were significantly associated with APP and SPLT scores
averaged over all environments.

The nine markers significantly associated with canning quality traits in Population
1 individually accounted for 2.5 to 8.3% of the variation in APP and 2.9 to 9.9% in SPLT
(Table 29). Marker OY7.850 accounted for the highest amount of variation (8.2%) in
APP averaged over environments, followed by OP15.1150 (6.9%) and OPAGIO. 1650
(6.9%). Markers 0014.950 (8.4%) and OAGlO.1650 (8.4%) accounted for the highest
amount of variation in SPLT averaged over environments, followed by OP15.1150
(8.3%). The two markers in linkage group M2-l accounted for the lowest amounts of
variation in both APP and SPLT scores averaged over environments. Marker
OAN116.3000 accounted for 2.6% of the variation in APP and 2.7% of the variation in
SPLT, averaged over environments. Marker OGl7.1300 accounted for 3.6% and 3.5% of
the variation in APP and SPLT, respectively, averaged over environments (Table 29).

Tables 30 and 31 show the mean scores for APP and SPLT of all RILs in
Population 1 with either the marker band present or absent. For the seven markers in
linkage group Ml-l (OY7.850, 0014.950, OP15.1150, OAGlO.1650, OA17.4000,
018.1600 and OU20.1150), the allele associated with desirable APP and SPLT scores
came from MCM, the parent chosen in the study for its desirable canning quality. For the
two significant markers in linkage group M2-1 (OGl7.1300 and 0AN16.3000), the
alleles associated with desirable canning quality traits were derived from CDRK 82. For
each marker, the genotype of RIL 118-90, the highest scoring RIL in all environments,
was consistent with the allele associated with high APP and SPLT scores, whether the

allele came from MCM or CDRK 82 (Tables 30 and 31).

103

.25. 22. E “2:835 :08 E 208

 

 

 

6233 we 03.5 v5 988 se.—am: 55, Eva.— .. .36 u 8:85.36 we .33 3 E3556 3: 83 5x5: - m: a
3:25.835 5m 05 56 towns: - .m>=m =<e «53.5 mm been hoe—.5: - At 353a «mag hats: .. 3 6.32 ..
ed ad m: m: m: m: + Na MIA—U +
dd 9 dd 20.2 - Send—720
vd ad ad 8 9. dd 8 + da Ida—D +
dd ed nd ad 202 - 82.5—va
nvd ad t 202 t on ~ _.odDO
N.M m: V.M m: W.M m: w.N NW ECU I
9d ad ad dd + 202 + cow—.amo
N.M m: W.M m—u W.M w: G.N N” EMA—U I
vd ad ad ad + 202 + coovd 35
dd 3 ed a: ad 8 ad da MMQU +
ad ad ad ad t :02 t 82 .o_O<O
dd an ad 3 ad 3 ad Na :90 +
ad ad ad vd . 202 t on — fin 50
dd 3 ed 2. ad m: ad Na MMDU +
an an an an - Eu: - 83.50
_d 8 ad 3 ad 9. ad da MMOU +
vd ad ad ad t SUE t cmad>0
diam =< aaa— 0392 aaa— 5:2 aaa~ new: haa_ fix: aa— :52 baa— :omwz Ayala—Co 823
Meg do.“ 808 oleo>< 25°50 Me 3.58 .222 hoe—32

 

.93: .5. e5 AS aceeg
BM #30 350330 28 530536 3:83 05 mo 8528a 05 v5 .83.: 805 55 33688 academia? p.858 05
.3 8.0:“ Sienna 05 3: 35 a _ noun—30m .«o 3.8% 93 me 982 @8883 «e 35.33% he 8.58 cap—25. dd 038.

104

cab 22— 5 “5:88:60 :80 E 0.88 “mo—RE

 

 

 

 

a0 2:? Ba 208 $232 55 Beam e .36 u ounce—aim? ae .32 3 aganamm 8: $3 .8qu - m: a
85:89.35 x6 05 .85 aoafiofi . .m>E<e acumen mm was .5th - A+v 3:8? mm was poo—.38 - 3 8.0.2 n
Nd ad a: a: m: m: m: + «a via—U +
ad hd .202 - Sada—720
dd ad 8 m: m: ad 3 + Na via—U +
ad ad ad 202 - 82.200
dd ad .. EOE .. a2 _.aNDO
ad 2 vd an ad m: ad «a van—U -
dd rd pd ad + 202 + 82.20
ad 8 vd 2. ed 8 ed «a van—U -
dd ad ad ad + SUE + aged—<0
ad 2 ad 3 ed 8 ad da 0950 +
dd ad ad —d t 202 .. ado—.a_D<O
ad 8 vd 3 ad an ad «a Man—U +
dd ad ad —d t 202 .. a2 fid—ao
ad m: dd m: dd 8 ad ma 0550 +
an an 3. an - 202 - 8350
ad m: dd dd ad 2. ad da van—U +
dd ad ad ad ad t 202 t adafiwo
c.>=m =< aaS v:52 3.: :05 aaa. :22 82 :22 33 Sfiz aaS :22 8.8 _ ac 30:.“
and 8a 808 odmdu>< 259.80 .3 350a .222 5x32

 

.aa-a: 0:: 28 3:83 05 we mambo—Sm 2: v5 .335 805 5a? @238me €505:me £8.85 05 ac
3.2—a 0:895 05 an: 85 05— a zeta—.58 ao 3.5% 92 ac £53 @8885 ac 9558 ac gov 8a 8.58 oafio>< .: 05$.

105

The marker composites were not significantly associated with APP and SPLT in
all the environments and reflected the environmental specificity of the individual markers
(Table 32). The composites accounted for about 5 to 21% of the variation observed in
both APP and SPLT. Marker composite A - all 11 markers, in both linkage groups M1
and M2 - accounted for the greatest amounts of variation in both APP (14 to 21%) and
SPLT (16 to 21%), followed by marker composite B - linkage group M1 (11 to 13% for
APP and 12 to 14% for SPLT). Marker composite A (all markers) was significantly
associated with APP and SPLT in Michigan in all four years of the study, but was not
significantly associated with the traits in either Minn-1996 or NDak-1999. Marker
composite B (M1 markers) was significantly associated with APP and SPLT in Michigan
in 1996, 1997 and 1999, but not in Minn-1996, Mich-1998, or NDak-1999. The markers
in linkage group M2 (composite C) together accounted for 7.4% of the variation in APP
and 8.2% of the variation in SPLT (Table 32). This composite was significantly
associated with APP in NDak-1999, and with SPLT in Minn-1996 and NDak-1999.

The composite of one marker from each linkage group (marker composite D:
OP15.1150 and 0617 .1300) was significantly associated with APP and SPLT in five of
the six environments, and accounted for about 6 to 11% of the variation in APP, and 7 to
13% in SPLT (Table 32). The different flanking markers in Ml-l (Population 1) -
composite E: OY7.850 and OU20.1150 - and M1-2 (Population 2) - composite F:
OY.850 and OI8.1600 - accounted for about the same amounts of variation (5 to 9% for
APP and SPLT), and were significant in the same environments. The flanking markers of
linkage group M2 - composite G: OAH17.700 and OH18.1000 - were not significantly

associated with canning quality traits in any environment.

106

 

Table 32. Coefficients of determination (R2) for composites of RAPD markers

significantly associated wth scores for appearance and degree of splitting of processed
beans of 75 RILs of Population 1, planted in Michigan, Minnesota, and North Dakota

from 1996 to 1999.

 

 

 

Trait and Marker Commsites “b
Environment A B C D E F
Appearing
Mich 1996 0.170 0.125 - 0.089 0.071 0.087
Minn 1996 - - - 0.059 - -
Mich 1997 0.144 0.1 15 - 0.071 0.051 0.059
Mich 1998 0. 1 72 - - - — -
Mich 1999 0.176 0.1 13 - 0.098 0.088 0.090
NDak 1999 - - 0.074 0.061 - -
Overallc 0.21 1 0.132 0.080 0.1 13 0.082 0.089
Dem of splitting
Mich 1996 0.178 0.124 - 0.1 14 0.087 0.101
Minn 1996 - - 0.075 0.082 - -
Mich 1997 0.156 0.134 - 0.069 0.045 0.052
Mich 1998 0.174 - - - - -
Mich 1999 0.170 0.123 - 0.100 0.076 0.082
NDak 1999 - - 0.082 0.077 - -
Overall° 0.212 0.136 0.074 0.126 0.075 0.085
'Marker composites:

A - all markers (OY7.850, OQl4.950, OP15.1150, OAG19.650, 0A17.4000,

018.1600, OU20.1150, OAHl7.700, OGl7.1300, OAN16.3000 and 0H18.1000).
B - markers in linkage group Ml (OY7.850, OQ14.950, OP15.1150, OAGlO.1650,
OAl7.4000, 018.1600, and OU20.1150)

C - markers in linkage group M2 (OAH17.700, OGl7.1300, OAN16.3000 and
0H18.1000)

D - one marker each from MI and M2 (OP15.1150 and OGl7.1300)

E - flanking markers from linkage group Ml-l (OY7.850 and OU20.1150)

F - flanking markers from linkage group M1-2 (OY7.850 and OI8.1600)

G - flanking markers from linkage group M2 (OAH17.7OO and 0H18.1000)

b (-) - not significant at level of significance = 0.05.
c Overall - APP and SPLT scores averaged over all environments

107

The marker groups were used to select RILs from Population 1 to test their ability
to select lines with desirable canning quality (Table 33). Seven lines had the 11 marker
alleles associated with desirable canning quality and were selected using marker
composite A. The average scores for APP and SPLT of these selected lines were 3.5 and
3.3, respectively, averaged over all environments. For comparison, the population means
were 3.3 and 3.1 for APP and SPLT, respectively. The individual lines selected using
composite A are shown in Table 34. Selection using the markers in linkage group M1

(composite B) resulted in 30 lines, with an average score of 3.4 for APP and 3.3 for
SPLT, averaged over environments (not shown). Selection with the markers in linkage
group M2 (composite C) resulted in fewer lines (17), which had lower average scores for
APP (3.4) and SPLT (3.2), averaged over environments (not shown).

Using one marker each from the two linkage groups (composite D) - OP15.1150
fi'om M1, and OGl7.1300 from M2 - resulted in 12 selected RILs (Table 33). These
selections had average scores for APP and SPLT of 3.7 and 3.6, respectively, which were
higher than the population means. The individual lines and their APP and SPLT scores
are shown in Table 34. Selection using the flanking markers in M1 (composites E and F)
resulted in a similar number of lines, 31 for composite E and 33 for composite F
(individual lines not shown). The average APP and SPLT scores of these selections were
equal to or higher than the population mean by 0.1 unit. Selection using the flanking
markers in linkage group M2 resulted in 18 selected RILs, which had average APP and
SPLT scores equal to the population means (individual lines not shown).

Selection using group M1+Gl7 (linkage group M1 and marker OGl7.1300)

resulted in 11 RILs, which had average APP and SPLT scores of 3.7 and 3.6, respectively

108

25:83.60 am 8.6 aoaabé 828m .Saa a5 mm< - =8o>0 a

88. 2:0 Ea ©3539 N2 95a omfié 58 Ease mafia: - o
88. a5 23 5.59 22 use» vase: 58 333:. ways: - a
82 380 as 02.59 E: gem owes. :3 £9.35 3an - m
89: .too as o: 3:9 2 Ba E 58 £8 338 as - 9
d2 nap—a gas—c: E who—BE - U
.2 95% mas—5. £ who—BE - m

mane—.88 =a - <
“anamoafioo 5:32.

 

 

 

 

 

 

 

ad ad :d dd dd ad dd dd dd _d dd ad samba/O
dd dd ad ad ad dd ad ad :d ad - ad aaa: V:52
ad ad ad ad ad ad ad ad ad ad ad ad aaa: :02
ad ad ad ad ad ad ad ad dd ad ad dd aaa. :92
ad ad dd ad ad ad ad ad ad ad ad ad 32 52
dd _d ad ad ad _.d ad ad ad ad - dd aaa_ :2
ad ad ad ad _d ad ad _d dd ad 2 dd aaa_ :um2
32 auto—um was 3:88 ao magnum ao uoamoa 8a 808 :82
ad ad dd dd dd ad dd dd dd dd dd ad E395
dd :d ad ad ad :d ad ad ad ad - ad aaa_ 8152
_d ad ad ad ad :d ad ad ad ad ad :d aaa_ :02
ad ad ad ad ad ad ad ad dd ad dd dd aaa_ :22
:d _d ad ad ad :d ad ad ad ad .nd ad aaa_ :92
ad ad ad dd dd ad ad dd dd :d - ad aaa: 2
ad ad dd dd dd ad dd dd ad :d ad ad aaa: :u2
32 auto—um 93 3:23 ao 3530an 8a 208 :82

e: :a a: 43 aa aa pa ea E S

a _ Z<+ _ 2 a _ Di 2 O m m D U m < :88 Mano 2U2 2685525

33% cog—«aca— .83.—

 

dozmqfiaou 5:82-30» Em BSA—mam 38:38
.8 858a mam: auto—om 32 _ nous—=99— .a0 88: @3885 ac madam—mm .ao oohaoa a5 3532:? 8a 888 :32 .dd 2:5

109

(Table 33). The individual RILs selected are shown in Table 34. Selection using group

M1+AN l6 (linkage group M1 and marker OAN 16.3000) resulted in 10 RILs, which had
average APP and SPLT scores of 3.8 and 3.6, respectively. These 10 RILs are shown in
Table 34. Both groups M1+Gl 7 and M1+AN16 resulted in RILs with APP and SPLT

scores higher than the population means for these traits.

Table 34. Average scores for appearance and degree of splitting of processed beans of RILs of

Population 1 that were selected using marker composites composites A and D, and Ml+Gl7

(linkage grog) M1 and OGl7.1300), and M1+AN 16 linkagggroup M1 and OAN16.3000).
Trait and overall

 

 

Accession scores‘ Marker Composite
APP SPLT A D M1+Gl7 M1+AN16

1 18-05 3.9 3.6 "' "' *
1 18-09 3.8 3.7 * "' *
1 18-21 3.7 3.6 * * "‘
1 18-22 3.4 3.2 * " *
1 18-42 3.2 3.1 " "‘ *
1 18-49 3.4 3.2 * "
1 18-63 3.4 3.3 "'
1 18-72 3.6 3.4 * " *
1 18-81 2.8 2.8 "‘ * "'
1 18-90 5.8 5.9 "' "
1 18-94 3.4 3.3 " “ "'
1 18-95 3.9 3.6 "' "' "'
Population Mean 3.3 3.1

 

' Overall Scores -averaged over environments

Based on the number of lines selected, and the APP and SPLT scores of these
selected RILs, the best marker subsets for identifying RILs with desirable canning quality
were D, Ml+Gl7, and Ml+AN 16. These subsets permitted the selection of 12 (average

APP = 3.7), 11 (average APP = 3.7), and 10 lines (average APP = 3.8) with dasirable

110

canning quality, respectively (Table 33). The lines selected using these sets of markers
had average scores higher than the population means for both APP and SPLT in all
environments. Seven RILs were common to the groups of RILs selected using these 3
marker subsets (Table 34). Marker subsets Ml+Gl 7 and M1+AN 16 permitted the
selection of an RIL with less than desirable canning quality - RIL 118-81, which had an
average APP and SPLT score of 2.8 (Table 34). This RIL was not selected using marker
composite D. Marker subsets M1+Gl 7 and M1+AN 16 also permitted the selection of 2
RILs, which had desirable canning quality scores but which were not selected using
marker composite D. These 2 RILs were 118-72 (average APP = 3.6) and 118-95
(average APP = 3.9). The three marker subsets - D, M1+G17 and M1+AN 16 - permitted
the selection of RIL 118-90, which had consistently desirable canning quality traits across
environments (average APP = 5.8).

Verification of putative RAPD markers in Population 2. In Population 2, only the
markers in linkage group M2 - OAH17.700, OGl7.1300, OAN16.3000 and OH18.1000 -
were significantly associated with APP and SPLT (Table 35). Two markers -
0AH17.7OO and OH18.1000 - were not significantly associated with APP and SPLT in
Population 1 (Table 29), but were significantly associated with these traits in Population 2
(Table 34). In general, the markers in linkage group M2 accounted for larger amounts of
variation in Population 2 than any marker in Population 1.

Marker 061 7.1300 was significantly associated with APP and SPLT in Minn-
1996 and NDak-1999. In NDak-1999, OGl7.1300 accounted for about 14% of the
variation in APP and SPLT, the largest amounts of variation accounted for by any

individual marker in both populations. Markers OAN16.3000 and OH18.1000 were

111

significantly associated with APP and SPLT only in NDak-1999. Marker OAN16.3000
accounted for about 6% of the variation in APP and SPLT, while marker OH18.1000
accounted for 7 to 8% of the variation in the traits. Markers OGl7.1300, OAN16.3000
and OH18.1000 were significantly associated with APP and SPLT scores, averaged over

all environments, and accounted for 3 to 4% of the variation in these traits.

Table 35. Coefficients of determination (R2) for RAPD markers in linkage group M2 that
were significantly associated with scores for appearance and degree of splitting of 73 RILs
of Population 2, planted in Michigan, Minnesota, and North Dakota from 1996 to 1999'.

 

 

 

 

Trait and Environment
M2 Markersb Mich Minn Mich Mich Mich NDak All Env.‘
1996 1996 1997 1998 1999 1999
Values of R2
Amce (APP)
OAH17.700 - - - - - 0.097 -
OGl7.1300 - 0.041 - - - 0.140 0.044
OAN16.3000 - - - - - 0.060 0.033
OH18.1000 - - - - - 0.081 0.032
Degr_ee of §plitting (§PLT)
OAH17.700 - 0.044 - - - 0.097 -
OGl7.1300 - 0.046 - - - 0.142 0.042
OAN16.3000 - - - - - 0.063 0.036
OH18.1000 - - - - - 0.067 0.029

 

' M1 markers were not significantly associated with APP and SPLT in Population 2.
b (-) - not significant at level of significance = 0.05.
c Overall - APP and SPLT averaged over all environments

For the markers significantly associated with canning quality traits in Population
2, the alleles associated with high APP and SPLT scores were derived from CELRK

(Table 36). The difference in the average APP and SPLT scores of the RILs with either

112

95:89.35 £35 .8>o coma?“ - 5.55..

do... u 353.3% .6 .o>o. an Enema... .2. 83 €8.52: .. m: c
.58.... m. 2.3 56.38 - At .83.? m. .65.. .8th - A... ”0.0.? c
.N 8.338.. 5 5.5 c5 a? 5.3 83.8.8 3885:»... .2. 2c; :2 9.2» case: 5 “.252 .

 

 

Mon ch mu.— m: mun mfl m: gun—mu I
an 3 202 + 8:350
dd md m: m: m: mu m: Mum—m0 +
~.m QM 202 . cane—EC
a... 3. a. E 8 3. a. 55.5 +
—.m QN 9N 202 . GOES—00
3. 8 a. a. an 8 53.6 +
m: od a.N 202 - OAKS—EC
watt—mm mo gun
and md m: m: m: m: m: 21de ..
N m QN 202 + coo—.w—IO
mnn «an m: m..— m... m: m: vans—EU + .
N m o. m 202 . coon c _Z<O
Wm WM w: m: m: WM 2. v—M—x—HU +
N.m w.N QM 202 u com—.a—OO
mnn m: m: m: m: m: v—M—AHU + .
m: a N 202 . 8a a .EO
oo§mm<
L355 cam—dag $37.20.: mam—icon). baa—ABSZ cammiqcmz cam—23.2 0.0.3
.c 83cm .222 .352 2a a...

 

 

.985. 3. 2% #5 258.5
23 8.8.52 553.. 2.. a6 8305» 2.. Ba .55 owe... 5.? 33.088 3.52.256 88.38 2.. ..o 8.0.? 2.3%? 2...
.5. a... 35. d 5.3.2.9. .0 3:8» 93 we v.52. .oommooen. we m:.:..% do 33% c5 8:839? 8.. v.28» own—03‘ .ad 2%...

113

marker allele was as much as 0.7 units, in the case of the alleles of marker OGl7.1300,
for APP in NDak-1999.

Only the marker composites with markers from linkage group M2 - composites A,
C, D and G - had significant associations with APP and SPLT (Table 37), as expected
from the associations of the individual markers with the canning quality traits. Marker
composites B, E and F were not significant in any environment. Marker composite B was
composed of the markers in linkage group M1 while composites E and F were of the
flanking markers in M1. Marker composites A (all the markers together) and C (markers
in linkage group M2) - was significantly associated with APP and SPLT only in NDak-
1999. In this environment, marker composite A accounted for 22% and 24% of the
variation in APP and SPLT, respectively, while composite C accounted for about 16% of
the variation in APP and SPLT.

Marker composite D (OP15.1150 and OGl7.1300) was significantly associated
with APP and SPLT in Minn-1996 and NDak-1999, and with scores averaged over
environments (Table 37). This composite accounted for 5 to 17% of the variation in the
traits. Marker composite G (flanking markers in linkage group M2, CAI-117.700 and
0H18.1000) was significantly associated with APP in Mich-1998 and NDak-1999, and
with SPLT in Minn-1996 and NDak-1999.

When the composites were used to select RILs in Population 2 (Table 38), the
results were similar those in Population 1. Based on the number of RILs selected and the
average overall scores of the selected RILs, the best composites for selecting RILs with
desirable canning quality were composites D, Ml+Gl7 and M1+AN16. Marker

composite D resulted in the selection of 13 RILs, which had average overall scores of 3.6

114

Table 37. Coefficients of determination (R2) of composites for RAPD markers
significantly associated with scores for appearance and degree of splitting of processed
beans of 73 RILs of Population 2, planted in Michigan, Minnesota, and North Dakota
from 1996 to 1999.

 

Trait and
Environment

Appearance
Mich 1996

Minn 1996
Mich 1997
Mich 1998
Mich 1999
NDak 1999
Overallc

Degee of splitting
Mich 1996

Minn 1996
Mich 1997

A

0.223

Marker CompositesIla

C

D

0.050

0.174
0.053

0.057

G

0.043

0.121

0.047

 

Mich 1998
Mich 1999 - - - - -
NDak 1999 0.244 - 0.158 0.182 - - 0.119
Overallc - - - 0.056 - - -

 

“Marker composites:

A - all markers (OY7.850, OQl4.950, OP15.1150, OAGl9.650, OAl7.4000,
018.1600, OU20.1150, OAH17.700, OGl7.1300, OAN16.3000 and 0H18.1000).
B - markers in linkage group M1 (OY7.850, OQ14.950, OP15.1150, OAGlO.1650,
OA17.4000, 018.1600, and OU20.1150)

C - markers in linkage group M2 (OAH17.700, OGl7.1300, OAN16.3000 and
0H18.1000)

D - one marker each from MI and M2 (OP15.1150 and OGl7.1300)

E - flanking markers from linkage group Ml-l (OY7.850 and OU20.1150)

F - flanking markers from linkage group M1-2 (OY7.850 and OI8.1600)

G - flanking markers from linkage group M2 (OAH17.700 and 0H18.1000)

b (-) - not significant at level of significance = 0.05.

c Overall - APP and SPLT scores averaged over all environments

115

88......8 as. 83.58. N: .8.» 0.9.5. as. 5.58 mans... - o
88....8 Ea 8.50. «:2 .8... 83.5. so... 9...“... 8:5: - h.
82 .580 as. .350. 7:2 .8... 0.5.... so... 3.35 mags... - m
88:60 as. on. 3.8. «.2 as. .2 so... 58 5.2. 25 - a

N2 :8.» 888.. :. $8.88 - U

:2 9.2... 0.5.... a. £3.38 - m

 

 

 

 

 

. 88.8.: .... - <

m.:o.::o..>:o .8 .26 .8882. 888 Exam :8 55¢. - =85 .. ”8.62.88 8.83..
«an Om van N.m N.m 1m 0m N... M... N... 1m a.m =....8>O
0.n. w.m Em N.m N.m w.m Em N... N... ..m a... N.m aaa. 89,.
in Em 1m 1m 1m o... v... m6. 1m v... N... N.v aaa. 8.2
..m o... Wm ..m ..m o... v...” ..m a.N N... w.N a... waa. 8.2
.mm a... hm hm hm a... 9m w.m hm hm ad a... Na. 8....
..m Nd ..m a.N «N N... ed a.N m.n m.N N.N N... oaa. :52
Qm N.m w.N m.N w.N ..m EN w.N o6. w.N n.N in waa. no.2

.13.... 3.8.8 .8 gm... .8 com“... 8.. 08%
On 0.... Wm Wm m.m 0.n 1m 1m md m6. Em N... 3.8.30
9m m.m 0.n N.m N.m Em 0.n N.m N... ..m a... N... aaa. .752
Wm w.m Wm Wm m6. hm hm Wm 0.n in 0.n vé aaa. 8.2
..m ox.“ v... ..m ..m O...” in ..m w.N N.m .N a... waa. no.3.
w.m a.m 9m Em Em a... 9m wd w.m fin N.m a... Na. no.2
v... Wm v... N.m N.m v... Om N.m 1m N.m 9N wd ea. 8.2
m... w... in v... in fin m.m 1m Wm v... N... .6 oaa. no.2

ma... 8.8.3. .8 38.2312

8.. .2. .2. ac 8... .2. G: a... 5
o.Z<+..>. 20+... 0 .. m n. U m < :88 v2.50 202 80:82.8...
:3 8.8.2.0.. .8: .8...

 

.:o..a:.n:80 8.802.80ng 88.85 88.8.:
8 8.8.» m5... 8.8.9.. 3.... N 8.8.2.0.. 8 «:8: 8888... .8 w:........ .8 3%.... :5 3582...: 8.. $88 :32 .mm 0.8...

116

and 3.4 for APP and SPLT, respectively. Marker composite Ml+Gl 7 resulted in 12
selected RILs, with average overall scores of 3.6 and 3.5 for APP and SPLT, respectively.
Marker composite M1+AN 16 resulted in 10 selections, with average overall scores of 3.5
and 3.3 for APP and SPLT, respectively. The average overall scores for the RILs selected
using these marker compositas were higher than the population means for APP (3.3) and
SPLT (3.2).

The RILs selected using composites A, D, M1+Gl7 and M1+AN 16 are shown in
Table 39. Nine RILs were commonly selected using D, Ml+Gl7 and M1+AN16. Two
of these, RIL 119-45 and RIL 119-94 had less than desirable APP and SPLT scores. The
other RILs, including those not commonly selected using these composites, had desirable
APP and SPLT scores. Composites D and M1+Gl 7 permitted the selection of RILs with

desirable canning quality traits, which were not selected using composite M1+AN16.

117

Table 39. Average scores for appearance and degree of splitting of processed beans
selected using marker composites A and D, and M1+Gl 7 (linkage group M1 and
OGl7.1300), and M1+AN16 (linkagggroup M1 and OAN16.3000).

 

 

Accession Trait and Marker Composite
Overall Scores‘| A D Ml+Gl7 M1+AN16
APP SPLT
1 19-14 3.8 3.7 " *
1 19-36 2.8 2.6 "'
1 19-42 3.6 3.6 * "‘ *
1 19-45 2.9 2.7 "‘
1 19-53 2.5 2.3 "‘ "' * *
1 19-54 3.4 3.2 * " "‘
119-55 3.3 3.2 * *
1 19-64 3.2 3.1 "' "' "' “
1 19-65 3.7 3.4 "' * *
1 19-69 4.5 4.0 * “ * “
1 19-71 3.7 3.3 * "‘ "
1 19-72 4.5 4.5 "‘ *
1 19-78 4.6 4.6 "' * "‘ "'
1 19-94 2.8 2.6 " * * "‘

 

' Overall Scores -averaged over environments

118

DISCUSSION

Breeding programs for kidney beans and other dry bean market classes must
consider consumer preferences for canning quality, in addition to yield and other
agronomic characteristics. Although desirable canning quality traits are important, bean
producers do not accept varieties solely for these characteristics. In the same way, high-
yielding breeding lines are also subject to the strict requirements for canning quality
sought by processors and consumers. Thus, in addition to the yield potential of RILs
comprising Populations l and 2, which were evaluated in this study, canning quality of
these RILs must be evaluated concurrently.

The significant negative correlations detected in the two populations, between the
canning quality traits (APP and SPLT) and seed weight complicate breeding for these
traits simultaneously. These findings are in agreement with Fomey et al. (1990) who
found significant correlations in kidney beans between large seeds and splitting during
processing. The negative correlations observed among canning quality traits, yield and
seed size have important implications in breeding for these traits. Breeders must devise
strategies to select for these traits simultaneously. Considering each trait separately may
lead to the improvement of one at the expense of the other two. In breeding for canning
quality, for example, the breeder must ensure that seed size and quality are not altered
beyond acceptable limits. Since seed size in dry bean is subject to a federal grade
restriction by the USDA Agricultural Marketing Service, seed size for the market class is

an important criterion for the acceptance or rejection of a cultivar by the bean industry.

“9

Canning quality of two kidney bean recombinant inbred populations

Although several factors influencing the appearance of canned beans - such as
brine clarity and the amount of starch in the brine and on the surface of beans, the degree
of clumping, and seed color, size and shape - are considered in evaluating the
acceptability of canned kidney beans, the degree of splitting of the beans also plays a
large role in acceptance or rejection of a sample (Lu and Chang, 1996; Fomey et al.,
1990). The causes of splitting during processing are not known, although factors such as
genotype, condition of the seed at harvest, storage practices and processing methods may
affect the trait. The positive correlations between HC- a measure of how well a bean
hydrates during soaking — and APP and SPLT indicate the importance of factors that
affect water imbibition during processing. These factors, which include the physiology
of the seed coat and cotyledons, affect the degree of splitting of the beans and thus overal
quality. The high positive correlations between APP and SPLT indicate that a single
rating will suffice to evaluate the canning quality of kidney bean lines. The use of APP
alone will include perceptions of the degree of splitting of the beans and other traits that
affect the appearance of the processed beans.

In both Populations l and 2, first-order interactions (genotype x year and
genotype x environment) were significant for APP and SPLT, indicating that either a
change in the ranking of the genotypes or the degree of the differences between them
occurred. Some RILs in both populations had consistently desirable canning quality
across years and year-location combinations (environments). The significance of these
interactions may therefore indicate a change in the degree of differences among the

genotypes rather than a possibility that the performance of a line may differ drastically

120

from one environment to another. Moderate to high estimates of heritability for APP and
SPLT indicate that selection should be effective in developing lines with desirable
canning quality. Walters et a1. (1997) reported heritability estimates in navy bean of 0.48
to 0.78 for the components of canning quality - visual appeal (VIS), texture (TXT) and
washed drained mass (WDWT).

The significant effects of the environment on canning quality traits in this study
were expected, based on experiments conducted by others over a span of 20 years
(Ghaderi et al., 1980; Hosfield and Uebersax, 1990; Hosfield et al., 1984b; Walters et al.,
1997). What allometrically correlated plant characteristics and developmental aspects of
the seed that accounted for variable canning quality responses to a range of physical
environments are present unknown. However, the results from studies in grain crops
(Borojevic and Williams, 1982; Wych et al., 1982) suggested that stresses induced by the
environment during seed development had large effects on seed characteristics. In the
Michigan environments, seasonal temperature and rainfall patterns (not shown)
conducive to a stress environment prevailed at times during the seed development period
in some of the years during which the experiment was conducted.

Location effects including the presence of foliar pathogens on the experimental
materials may also have contributed to stress influences on seed development. The
climate prevailing at harvest (not shown) is an additional factor that may have influenced
the results of this study. Cool and wet climatic conditions at harvest may also have a
negative effect on canning quality. Seeds with high moisture content at harvest are often
discolored and sprouted and must be removed to ensure that the sample falls into a

marketable grade. Moreover, marketable beans high in moisture have high respiration

121

rates (Gonzales et al., 1982; Kays et al., 1980). High respiration rates of beans indicate
that major metabolic pathways in the seeds are activated (Kays et al., 1980), which could
lead to physico-chemical changes affecting canning quality.

Sufficient variability in canning quality was observed among RILs in both
populations to select individual RILs that scored as high as, or better than, commercially
accepted varieties used as checks for canning quality traits. The RILs, 118-90 and 119-
34, and the others with consistently desirable APP scores show potential for use as
sources of genes for desirable canning quality. However, these two RILs had
significantly lower yields than the checks. RIL 118-90 had an average yield of 1657
kgoha'l over the six environments, and was among the lowest yielding entries (Appendix
Table A.1). Line 119-34 was likewise one of the lowest yielding lines, with a yield of
1983 kg-ha'l averaged over all environments (Appendix Table AB). The seed weights of
both lines were within the range of 50-60 per 100 seeds (Appendix Tables 2 and 4),
which is considered acceptable for the kidney bean (Adams and Bedford, 1975), although
line 118-9O is at the low end of this range. The low yields of these two RILs, which both
had desirable canning quality, offer further evidence for the negative correlation between
yield and APP. Although RIL 118-90 and RIL 119-34 are both low yielding, these lines
merit consideration for further testing in a dry bean breeding program. In crosses with
high yielding genotypes, one would strive to “capture” the genes for desirable canning
quality carried by these RILs, and combine them with genes for yield from the high-
yielding parents.

Deshpande et a1. (1984) foresaw an increased demand for canned beans, due to

their availability as easy-to-prepare or ready-to-eat food. This demand is not likely to

122

decrease, with the increased interest in convenience and fast food. The contribution of
beans to human nutrition as an alternative protein source will also serve to encourage its
consumption, especially among those who choose a vegetarian diet. An increased
demand for kidney bean varieties with desirable canning quality should follow an
increased consumption of beans of this market class. The identification of genotypes
with superior canning quality represents continuing efforts to meet these demands.
Improved understanding of the inheritance of the trait and the effect of the environment
and genotype x environment interactions serve to broaden the information on which
breeding approaches for the trait are based.
RAPD markers for canning quality traits in kidney bean

An important factor in applying selective genotyping in identifying markers is the
choice of lines to include in the bulks. The composition of the bulks must be such that the
extreme phenotypes are represented and that the bulks differ as much as possible only in
the genomic region or regions of interest. For example, in the case of the bean traits
under consideration in this study, five RILs were selected based on their desirable APP
and SPLT scores in four environments. Five other RILs were selected to comprise the
bulk with undesirable canning quality traits. The choice of RILs for the bulks is
especially important when the contrasting phenotypes differ by degrees, as in this study.
Unlike single-gene traits, quantitative traits cannot be categorized into discrete groups
that differ markedly in their phenotypes. Variation is continuous, with slight differences
between genotypes commonly indistinguishable. Significant environmental effects also
cause the differences between genotypes to differ from one environment to another.

Thus, in addition to the performance of a line relative to the others, the consistency of

123

that line’s performance across environments should be considered. The number of lines
included in the bulk is also an important consideration and is affected by the size of the
population under study. A balance must be achieved between having too many and too
few lines in the bulks. Including too many lines in each bulk will increase the similarity
between the bulked DNA, thus making it harder to identify genomic regions that
distinguish one from the other. Using too few lines will increase the number of genomic
regions in which the bulks differ; some of these regions may not be at all involved in the
trait of interest.

In this study, 5 lines from Population 1, which comprised of 73 RILs, were
deemed as just the right number of lines for each bulk. RILs with APP and SPLT scores
at both ends of the 7-point scale, representing the most and least desirable phenotypes,
were identified in each of the four environments. The consistency of the lines’ scores
across environments was considered. The quality of the canned beans of the lines was
verified visually. In this last step, such as in the evaluation of most quantitative traits,
some degree of judgment was left to personal discretion regarding the choice of the lines.
The choices for the two bulks used in this study were thus based on three criteria: APP
and SPLT scores in four environments, consistency of the scores across these
environments and visual evaluation.

Eleven markers, which as a group accounted for a moderate but significant
amount of variation, were identified in the two kidney bean populations. The low
number of markers identified for canning quality traits may be due to several factors,
such as low levels of polymorphism between the parental lines detected by the RAPD

markers, and the incidence of many QTL with small effects on the trait which could not

124

be detected. Interactions between genes (epistasis), and genotype x environment
interactions (Paterson et al., 1991), which were significant in this study, may also be
contributing factors. Another primary obstacle in marker identification, particularly for
quantitative traits, is the precision of phenotypic evaluation (Luby and Shaw'), which in
turn is affected by the degree to which the genetic effects are confounded by genotype x
environment interactions. Unless care is taken in assessing the phenotype and ensuring
sufficient population sizes, the observed degree of correlation between the trait and a
marker in any population may not accurately reflect the actual degree of linkage. This, in
turn, ultimately affects the genetic efficiency of using MAS to improve the trait (Luby
and Shaw').

The eleven markers, in two linkage groups, associated with canning quality traits
indicate the presence of at least two QTLs, which influence canning quality traits in
kidney bean. The first linkage group, Ml (OY7.850, OQ14.950, OP15.1150,
OAGlO.1650, OAl7.4000, 018.1600 and OU20.1150), was putatively located on linkage
group B8 in the core map for P. vulgaris (Freyre et al., 1998), based on the amplification
products of 018.1600/1500 and OU20.1150 (Figure 10). The exact orientation of this
linkage group on the core map cannot be determined because of the different location of
the marker U20.1150 in the maps generated from the two populations, Ml-l and M1-2
from Populations 1 and 2, respectively (Figures 11 and 12). Based on the order of the
markers in Population 1 (Ml-l) (Figure 11), the linkage group M1 is at the lower half of
linkage group B8 of the core map (Freyre et al., 1998). Based on the order of the markers

in Population 2 (Ml-2) (Figure 12), M1 is at the center portion of BS (Freyre et al.,

 

' Luby and Shaw. Unpublished manuscript. Does marker-assisted selection make dollars and sense in a
fruit breeding program?

125

1998). The difference in the order of the markers in the two maps may be due to a
translocation involving the segment with marker U20.1150, which occurred in one
population but not in the other, or simply heterogeneity within the lines. Genes of known
function that are located on B8 are lipoxygenase (Adam-Blondon et al., 1994) and
glutamine synthetase (Nodari et al., 1993). Additional markers and populations with a
higher degree of inbreeding are needed to precisely determine the location of the M1
linkage group on the B8 of the core map, and to map the location of linkage group M2.

Linkage group Ml-l also has a greater total distance and fewer markers clustered
together (no recombination) than Ml-2. These differences may be due to more points of
recombination in this region of the genome in Population 1 than in Population 2. Greater
similarities may exist between the parents of Population 1 (both dark red kidney bean)
than between those of Population 2 (one light red and one dark red kidney bean), thus
facilitating greater crossing over at this region.

QTL involved in canning quality may be within or near the region of the kidney
bean genome represented by linkage group Ml. Except for marker U20.1150, the
position of a particular marker on the linkage group does not seem to affect the amount of
variation associated with that marker (Table 29). The QTL may thus be within the M1
region, but since the physical map distance covered by M1 is not known, the possibility
that the linked QTL resides outside the region cannot be excluded. Furthermore, if more
than one QTL is located at this region, one (or more) QTL may be within M1 and others
may be near the region. In either case, the QTL appear to be more adjacent to the
markers other than U20.1150, which accounted for less variation than the other markers

and was significantly associated with APP and SPLT in fewer environments (Table 29).

126

 

If QTL reside within the region of the M1 marker group, then the markers in the
center of the map have a greater potential for use in marker-assisted selection (MAS) for
canning quality traits, than the flanking markers. A map distance of 10 cM or less
between a marker and the QTL of interest aids in increasing gain from selection using
that marker (Paran et al., 1991; Kennard et al., 1994; Tirnrnerman et al., 1994). A marker
that is closely linked to the QTL or gene of interest facilitates faster improvement of the
trait during MAS than one in which there is a high degree of recombination with the

gene. The three markers at the center of M1, OP15.1150, OAGlO.1650, and OA17.4000,

 

have a total map distance of less than 10 cM in both populations.

Since it is not known whether the QTL is located exactly at the position of the M1
markers, the genotype of the QTL cannot be known with certainty (Paterson et al., 1991).
However, the source of these marker alleles is MCM, and the probability that the
associated QTL is also derived fiorn MCM is high. The use of MCM as a source of
genes for the improvement of canning quality in kidney bean breeding programs seems
justified by the results of this study. These linked markers, as a group, are associated
with a significant amount of variation in the APP and SPLT traits in canned beans in
Population 1 - 11.3 to 13.6% (Table 32).

The M1 markers segregated in a similar fashion in Population 2, which was
derived from a cross between a light red (CELRK) and a dark red kidney bean (MCM).
Even though MCM was a common parent in the two populations and presumably the
source of genes for the desirable canning quality traits in Population 1, the M1 markers
did not account for any significant variation observed in Population 2. Walters et a1.

(1997) reported population-specific markers in navy beans. Some of the population-

127

specific markers reported in the Walters et a1. (1997) study were monomorphic in all but
one of the three populations of navy beans studied; hence, the non-significant effects in
the other two populations. The small population sizes evaluated by Walters et a1. (1997)
may have contributed to the population-specificity of the markers. In the current study,
the number of R115 in each population was more than two times that used in the Walters
et a1. (1997) study. The eight putative markers in Population 1 of this study were also
polymorphic in Population 2 and did not deviate significantly from a 1:1 segregation
ratio. Segregation and similar linkage phases between the marker and the QTL are
important if markers identified in one population are to be useful in another (Dudley,
1993).

The second group of markers, linkage group M2 (OAH17.1300, OGl7.1300,
OAN16.3000 and OH18.1000), were significantly associated with canning quality traits
in both the dark red and light red kidney bean populations, and were not derived from
MCM (30, 31 and 36). Genotypes with undesirable canning quality are generally not
considered as sources of genes for improving canning quality traits. However, other
studies have shown that poor performing genotypes such as wild crop relatives, which are
rarely used in the improvement of quantitative traits, may in fact be used to improve
quality. Examples are QTL from the phenotypically inferior wild rice relative, Oryza
rufipogon used to improve grain yield in cultivated rice (Xiao et al., 1998), and QTL
from unadapted tomato germplasm used to improve color and soluble solids in cultivated
tomato (Tanksley and Nelson, 1996).

In the current study, CDRK 82 and CELRK appeared to transmit the QTL

detected by the markers in linkage group M2, which were significant in Populations l

128

and 2. Considering that the markers in M1 were not significant in both populations, the
two undesirable canning parents (CDRK 82 and CELRK) would not be likely to have the
same alleles for desirable canning quality at the loci represented by M2. Instead, alleles,
which do not contribute to the acceptability of canned beans, may be present in MCM at
the loci represented by M2. In some environments, MCM may exhibit average or even
undesirable canning quality because of the expression of these alleles. The substitution
of these alleles with others, even those from a variety with generally undesirable canning
quality, as may have occurred in these two populations, may improve canning quality.
The loss of these MCM alleles for undesirable canning quality traits is more important
than the source of the substituted alleles in promoting desirable canning quality. This
finding may explain the results of the experiments in Michigan in 1997 and in North
Dakota in 1999 in which MCM had lower than expected scores for APP and SPLT of
canned beans. In fact, CELRK, which was expected to manifest low APP and SPLT
scores (~3.0), had higher scores for these traits than MCM in these environments.

The QTL linked to the M2 are apparently more sensitive to environmental effects
than are the markers in M1. In Populations 1 and 2, OGl7.1300 was significant in both
Minn-1996 and NDak-1999 for APP and SPLT but was significant only for APP in
Population 1 in Mich—1999. Also in Population 1, marker OAN16.3000 was significant
only for APP in Mich-1999 and for SPLT in NDak-1999. In Population 2, this marker
was significant for APP and SPLT in NDak-1999. The other two markers, OAH17.7OO
and OH18.1000, show similar patterns of environment-specificity.

Unlike at the M1 region, only one QTL for canning quality may exist at the region

represented by M2. The flanking markers, OAH17.700 and OH18.1000, were

129

significantly associated with the APP and SPLT only in Population 1. The markers at the
center, however, were significant in both populations. These central markers -
OGl7.1300 and OAN16.3000 - thus appear to have more potential for MAS than the
flanking markers.

Since the two populations used in this study had a common parent, and dark red
and light red kidney beans are closely related market classes (Ghaderi et al., 1982),
markers common to both populations were expected to be found. This result was
observed only for the M2 markers, which were derived fiorn the parent other than MCM.
The QTL detected by M2 markers in Population 1 (dark red x dark red) were also
important in Population 2 (dark red x light red) (Tables 29 and 35). For the M1 markers,
however, this result was not observed. Several hypotheses are possible to explain these
results. Although the same QTL may be present in light red kidney bean RILs at the M1
region, these QTL may not be as important in the regulation of camning quality in this
market class as are other regions of the genome, such as the M2 region. This conclusion
is supported by the usefulness of the M1 markers in selecting lines with desirable canning
quality in Population 2, even though the markers did not account for significant variation
in this population. Otlner regions of the genome may have larger effects on APP and
SPLT in light red kidney bean. Other marker systems that generate higher degees of
polymorphism than RAPD markers may be able to identify these QT L. Other dark red
kidney bean populations derived fi'om MCM should also be investigated with regards to
the segegation of M1 markers and their effects on canning quality traits in those

populations.

130

Another explanation may be that the genome of both the light red and dark red
kidney bean has not been fully characterized. Additional markers may indicate that the
M1 region may be similarly important in Population 2. Still another possible explanation
for the lack of significant associations between markers and QTL in Population 2 may be
epistatic interactions between QTL, which masks their effect on APP and SPLT of light
red kidney beans (Dudley, 1993). Greater variability of the data in Population may also
have an effect in the identification of markers in this population. Further research is
needed to verify these hypotheses.

The use of only one marker for each linkage goup may be all that’s required to
effectively select for desirable canning quality in these populations. Based on the results
of the selection experiments on Populations 1 and 2, and considering the efficiency of
using the least number of markers possible, the best set of markers to use appear to be
one marker fi'om each linkage goup, particularly markers OP15.1150 and OGl7.1300
(marker composite D). Marker OP15.1150 has a distance of 0.0 cM fi'om OAG10.9SO in
Population 1, and fiorn OQ14.950 and OU20.1150 in Population 2. Selecting for this
marker, therefore, increases the possibility that the other three markers will also be
selected, whether in Population 1 or Population 2. Marker OP15.1150 accounted for a
relatively large amount of variation for APP and SPLT in Population 1, indicating eitlner
close linkage with a QTL with minor effects on canning quality or more distant linkage
with a QTL, which has a large effect. Marker OG1 7.1300 accounted for the largest
amount of variation in APP and SPLT in either population - 14% in NDak-1999 in
Population 2 (Table 35). The use of these two markers to select for RILs with caruning

quality resulted in similar numbers of RILs with similar average APP and SPLT scores in

131

the two populations (Tables 33 and 38). These results indicate that the two markers were
equally effective in selecting for canning quality in the two populations.

The heritability of the trait under selection and the proportion of additive variance
explained by a marker are factors that affect the efficiency of using MAS in improving a
quantitative trait (Luby and Shawz). Miklas et a1. (1995) suggested that for a marker to
be usefinl, it should account for variation as much as or more than the heritability of the
trait. In this study, the heritabilities for APP and SPLT are more than the variation
accounted for by any of the markers individually or by any composite of markers. The
stability of marker-QTL associations over environments is also another factor to consider
(Miklas et al., 1995). Environmental-specificity was observed in some of the marker-
QTL associations. Although the markers reported here have been shown to be effective
in selecting for desirable canning quality in these two populations, their use in other
populations needs evaluation, particularly in populations in which MCM is not a parent.
In such populations, the markers identified in this study may be useful in indicating a
genotype’s potential for desirable camning quality, even before seed production, and in
reducing the number of crosses needed to evaluate the trait (Dudley, 1993). The markers
could also be useful in reducing the number of lines to be planted, harvested, and
evaluated using conventional canning methods, saving considerable time and resources.

The QTL detected in this study offer important insights into markers, which may
be useful in breeding for canning quality in kidney bean. For example, QTL for desirable
canning quality traits may be present in the genomes of varieties showing undesirable

canning quality. Further investigations using other DNA marker systems on both dark

 

2 Luby and Shaw. Unpublished manuscript. Does marker-assisted selection make dollars and sense in a
fruit breeding progam?

132

red and light red kidney beans may shed more light on the genes responsible for camning
quality in these two kidney bean market classes.

In addition to the QTL detected here, other QTL responsible for a larger amount
of variation may be present in other regions of the genome. Likewise, additional minor
genes, each with a small effect, may be present. Further investigations with other marker
systems, such as AF LPs, and using other approaches may be useful to detect other genes
influencing canning quality, further define the linkage map presented here, and fine-map

the location of the QTL relative to krnown genes or markers in linkage goup B8.

Additional markers associated with either a few QTL with major effects or of numerous
QTL with minor effects need to be identified if MAS for canning quality in beans is to be
feasible. Mapping strategies that utilize saturated maps of the bean genome might prove
productive in identifying these additional markers. The presence of a low level of
polymorphism due to a narrow genetic base is another concern in using RAPDs. Marker
systems that generate a higher degee of polymorphism than RAPDs may allow the

identification of additional loci that were not detected here.

133

APPENDIX

134

. . .Bflfiflfioo

 

 

88 «.8 .2. 88 8.. 88. a... 8-...
.88 8... m... 88 8.. 88 8... 2-...
88 8... 8.. .8. 8.. 8.8 8... 8....
8.8 .8. .... .8. .8 8... 8.. .7...
28 .8. .8. 88 E. 28 8.. 8-...
..8 .8. 8.. .8. - - - .28.... . .3... 8.. 8.5.
.8. 8... 8.. .8. .... .8. 8.. 8....
.8. .8. 88 2.. .8. .8. 8.. 8....
.8. 8.. .8. .8. 8.. 8... E. ..-.:
:8 .8. 8.. .8. .8. 8... .8. 8.-.:
8.8 8.. ..8 a... .8. .8. 8.. 8-...
88 - - - 88 - - ....-...E.§=ez. 8.8.
.8. .2. .8. .8. .8. .8. 88 ..-.:
a... - - - - - 8.... ..E.§=e2a..8a.\§.3. .283.
.8. 8.... .8. .8. 8.... .3. .8. 8....
.8. .8. .8. R8 .... .8. o... 8-...
8.. .8. 3.. 88 8.... 8.. .8. 8-...
8.. 8.. .8. .8. .8. .8. .8. ..-.:
8.. .2. a... z... .8. 8.. 8.. 8-...
8.. .... 8.. 8... 9... 8.. 8.. ea...
c... .8. .8. .... 8.. - .8. .82.... .88..
.8. .88 88 .8. 8.. :8 2.8 8-...
.2. .8. .... .8. a... .... .... ..-...
9:85 82-882 .878... .878: 82-5.2 82-5.2 82.8.: 22.88..

 

8088.35 :80 e. we... 20;
.aaa. .8. ma. .naaa. .eaa. :. 88.2. 5.82

as 3805.: 5.20.: a. 8.5.. .88.? V.88 .8 .. 8838.. .e 28.8 2.. as a... 8 .e .72.... 8.0... ...< a...

135

 

 

. . 60:52.8
BVN ham _ vow.” wacN NowN c _ 3 mNhN ecum— ~
N _ N mch chm VNbN hm: _ _ .N amaN v0.3 _
~N¢N :N_ emcm N—mm mooN 2: 33 n7”:
QNVN 83 Nawm cmoN Sw— awn. RBN 2-x:
amvN _Nw_ wme wmmN can. chN 3:” No.3—
3: wVNN cNhN 8NN =5N :3 98m c _ -m. _
omVN mc _ N NmoN VmaN SnN SmN :2 wc.m_ _
chN wmg mmNm hme 8— N “.an a: 3&— _
chVN 3.3 NVNM _ 5N wMNN g _ N SMN $3 _
chN S.»— wag NmNN amt agN wean wmé:
wwVN m. _ N 0N. m cmmN §N ovoN wacN 3.x. —
_NmN 82 Nmmm cNhN 82 mwhN bmaN nN.w:
~ mm N 22 aNNm achN omoN SmN mme 3.3 ~
Nm mN an. N m _ hm aveN $2 a: 8K 9.58282
ommN XSN mwaN wme 8MN SVN ”SN 5?»:
hwnN n.6— MMNm SVN «5N amMN whoN NNé— _
anN Nag camN ScN m: n :NN ammN 3-3 ~
ammN 3a— mmmm NNoN 32 och thN Kuw— _
cth ammN wcmm SVN $2 E: m $3 awé~ —
:hN 3.2 3mm £MN mna VmwN nNaN 3-»:
vwnN Nan 32” NQVN 8NN SNN wcmN «Ww— _
mwnN 2.x— cmmm a; newN mam. OEN 31w:
.amN EbN mama GEN aNN mmaN mmeN 2-x:
a=8u>0 a8 _ .quZ 8a _ £22 waa 75.: 3a _ Aug: 93 _ :52 03 ~ -522 2283.3

 

 

...vo..:u..oo :3...va .20; ._.< 03....

136

 

 

Sivuzcmucoo
was 5 a: EN :3 $2 2% m3:
3: $2 83 SE :2 80 «N3 «1:—
33 8a 23 82 a: we: 58 $42.
83 Na 3% 3% 8% e3 «SN 8.»:
2a 8a $2 $3 28 $2 :2 8.”:
a: 2: E: 88 SR 22 28 N3:
82 a: 8% 32 $8 $2 «a: :3:
8% NS :2 “EN 23 22 EN 8;:
£2 a? :2 ”SN 3% $2 8: .3:
$3 - - - 3mm - $2 .282: \ 2.559% ~85.
2am «Na 2% 33 a: 38 $2 “1: _
82 8: £8 £3 82 :2 cg NE:
83 5: «SN 83 $2 38 2Q mi:
:2 «8 28m $8 82 83 $3 3.»:
$3 38 Sen Nam on: $2 28 m3:
:2 23 £2 28 «RN an. as 3.”:
$8 N; :2 8mm 22 8: Ba 8.»:
82 SM: wan 2a 8: $3 88 2-»:
32 8a .5... 83 3 m 6: EN 3-”: ~
32 we: can 8: 2.: - 88 as“: 3.
MEN 88 2% 3: ME 88 :2 an»:
38 a: $2 EN EN as 88 3.»:
8mm 2.: 28 $3 ”a: 2: 8% 5m:
2826 2&7; §7§2 ”8322 82-55 08753 eavfiz .=o_a§<

 

35835 £8 a Ca. 20$
:33580 p.23: ”so; ._.< 2.3

137

mounts, «025 - a
noun—anon 05 .3 3:8!— - 9
«888.530 55 comp—o: - =55.—

Aagogao =a 83 vamp—03v 20$ :93 €83 8 €88 woman—.3 893 83808< .. .

 

o_o_
smu—
h_w_
www—
moon
m_om
~vo~
_mc~
when
vaom
macm
n¢_N
_N_N
aN_N
mm_N

“tw—
cam—

~hv
wan—
unc—
wNw
NVm
c—a
aa
Nom—
mm~_
~a~
m~w
emw
eaa
~mb
mwm

c;—
>m_n

amwm
mam—
oamm
vwmm
mth
m_MN
mhwm
_NwN
oan
~n_m
mhhm
~_Nm
vo_m
VNcN
moan

Num—
NooN

wm—N
hemm
www—
e_—N
aamm
_¢a_
cuou
oo—m
whvu
°¢m~
cawm
~mn~
amwN
mean
_hNN

Awa—
anN

c~c_
nae—
ace—
wow—
.m—m
«wan
m—nu
ccvm
owmm
«hem
«mm—
wea—
wmvm
¢_VN

mva
b¢-N

ana

awn
o¢-
emo—
ame—
cmm
mum—
numm
_ma
awn—
ve—m
man—
hae—
w~a~

fi—m
m—0N

—me_
mm—N
nw—m
warm
«ham
wvom
m_c~
wch
mmmm
aamm
wnmm
mwNN
mwmm
n-«m

3 8&5, Ho 203580
502

ewg _
onwa— _
Emm— _
8.3 _
Naé— _
3-3 _

on MM—QU
3.2 —

h 73 —
8-3 —
ecu—80 \ 33am vow—v membe—
.31w__
8.3 _
Q13 _
~m1w__

 

2326 82-; 32-55 @8222 §7§2 23753 08222
“5.58:8 €43 cleélgl. 23>

 

.uommmooot.

.. 6025.80 Arms—dye m2£> ._.< 05$.

138

 

 

. . .mvuach—GOO
E» ».»». ».»» ».»». ».»» ».»» ».»» _»-»:
E» S.» ».»» ».»» ».»» ».»» ».»» 31»:
E» «.2 ».»» ».»» ».»» ».»» ».»» 87»:
».»» ».~» ».8 ».»» .8 ».»» ».»» »»-»:
».»» v.3. ».»» ».8 .2» ».»» ».»» 8.»:
_.»» ».»» ».»» ».»» ».»» ».»» ».»» 8-»:
».»» ».»» ».»» 3» ».»» ».»» ».»» »»-»:
».»» v.2. ».»» ».»» E» ».»» «.8 :-»:
».»» ».»» ».»» ».»» ».»» _.»» ».8 3.»:
».»» S» 2.» ».»» ».»» ».»» ».»» 8.»:
».»» ».»». ».»» ».»» ».8 ».»» 3» »9»:
».»» ».»» ».»» ».»» 3» ».v» ».»» ».w». _
».»» ».»» 3» ».»» ».»» E» ».»» 2-»:
».»» ».»» ».»» ».»» ».»» ».»» ».»» a»:
».»» <3. ».»» ».»» ».»» ».»» «.8 ~»-»:
».»» ».»». ».»» ».8 ».»» ».8 ».8 S»:
».»» ».»» ».2 ».»» N8 - ».»» 93...: :8»:
e8 ».»» ».»» as» 9.2 ».»» «.8 3-»:
S» ».»» ».»» ».»». S» 3.» 8» R»:
E» 3» 2K ».8 ».»» ».»» ».8 8.»:
».»» - - - ».»» - - 93722382» 833.
».»» 3» 2» ».»» ».»» ».»» ».»» 2-»:
».»» 2» ».»» ».»» S.» ».3 ».»» 3-»:
L330 32.322 32.5% »8222 33.53 82-5»: 32-53 .8588...

“55:83.5 :80 E A730» 8:3 2203 teen

 

.82 Ea »2: .8»: .3»: a 59.5 532 E» 38053
.3222 5 35:. 5:05, .32: as» ._ 8388 »o 3:23 05 Ba 3:. »» »o 88» 72: .3 35:03 v8» .~.< 2.3

139

 

 

...B§C.5COO
».»» ».»». ».8 ».»» ».: »._» ».»» 8.»:
».»» ».»» ».»» ».»» ».»» “.3 . ».»» a»:
».»» ».»» ».3 E» ».»» ».»» ».»» . »»-»:
».»» ».S ».»» E» ».»» 2» _.»» .5382
».»» ».»» ».»» ».»» _.»» ».»» ».G 3.»:
».»» ».»»- _._» ».»» S.» ».»» ».»» aw»:
».»» ».»». NS ».»» 3..» ».»» ».»» 2-»:
».»» ».»». Z» 5.» G» ».»» 2» 8.»:
S» E» ».»» ».»» ».»» I» ».»» 8.»:
Z» ».»». ».»» ».»» q.» 3» Z.» 8-»:
E» :4 ».»» E» ».»» ».»» ».»» »»-»:
E» E». ».»» ».»» _.»» ».»» ».»» 3.»:
E» ».»» ».»» 2» ».8 ».»» ».»» »7»:
».»» ».»» 2..» Z» ».»» ».»» ».»» »9». _
».»» ».»» v.3 ».»» 3» E» ».»» 3-»:
».»» ».t- ».3 Z» 3» ».»» 3.» 8-»:
3» ».»» ».»» E» «S 3» ».»» 2.»:
E» ».»» ».»» ».8 ».»» ».»» ».»» NY»:
».»» ».3. ».»» ».»» ».»» ».»» «.8 a»:
».»» ».»»- ».S ».8 ».»» ».»» E.» 3.»:
3.» SS 2» «.8 3.» S.» ».»» 8.»:
».»» ».»». we. ».8 E» Z» 3» 3.»:
E» N:- _.»» ».8 Z» ».»». ».»» 2-»:
.2826 cavfinz 23222 »8222 23.-5:2 @3753 @8222 .5388»-

»§8$5 :08 5 r-voo» 8:3 Emma? voom I
. . 60::qu 38» 78:3 8:303 voom .N.< 0.n—ah

140

. . .vosczcoo

 

 

».»» as. ».«e 3‘» «.8 ».»» ».»» Q»:
».»» ».»» ».»» ».»» 9: 9% ».»». 3.»:
2.» 3»- ».«» ».»» q:- 34 ».»» :1»:
3.» _.«v «.8 ».»» «.3. 2»- ».«» u«» 550
3.» ».»». ».»» ».»» ».S ».»» ».»» 8-»:
3» ».»» S» E» ».»». 2.» ».»» ».»»:
».v» «.8 ».»» ».»» - - - .35 \ “:5: 86 82.3.
».»» Se ».»» ».»» I» «.».. 3» 8.»:
».v» ».«« 98 ».»» v.5 ».»» ».»» 2-»:
o.»» E». ».G 2.» v.«» _.»» ».»» 3-»:
v.»» as. ».»» «.8 c.3- 3» Z.» 3-»:
».»» ».3. ».G ».»» ».3 9% 3» 8.». _
».»» ».»». ».G ».»» v.8 ».»» ».»» 8.»:
».»» e. :- «. G ».3 - - - 23.38. \ UE»: use 8&8.
».»» «»v «.8 ».»» 0.8 3» ».»» a9». _
».»» v. G o.»» ».»» «i Z» _.«» «3. _
».»» ».«v «.8 «.»» «.8 ».»» ».»» »»-»:
».»» E «.8 ».»» v.8 »..«» ».»» 8.»:
».»» ».»». e. G ».»» ».»» 33 ».»» 8-». _
».»» ».»» ».»» ».»» 98 «.»» ».»» »»-»:
».»» ».3 3» ».»» ».»» ».»» ».»» 2-»:
S» «.3- ».8 ».»» ».8 a. _ » «a» «a». _
«.o» a».- E» ».»» ».2 ».»» ».»» 2-»:
.=§>o 8?; «8232 23222 «8322 087:5: e322: .8638».

Each—83:» :08 5 rboo» 8:3 3303 Bow I i
. . 63:95». €08 Tao—.3 3:903 voom .N.< nigh.

141

 

830....» ~86 - ..

83.3.8 22.. 285. - o

83%..» .26 town—oz. - Enos»

Egg...» =~ .26 coma-8.5V So.» :93 €33 8 9.5.88 coma—Ea 803 82808< - ..

 

 

«E a...» 8... .»..V «E «S Q... 83...; .o 25.0880
$3. «».»» .«.»» 8..» »«.»» »«.»» :82
.3» - - - - - to» 9.5.8.82238333. ««83.
«. . » n .v ».»» ».»» «.»» ».»» e...» 8-». .
»..» ».«» »..» «.»» ».8 «.»». ».»» 8-»:
»..» «.3. o.»» »..» «.»» ».«» ».«» »«-»:
»..» ».»» «.»» «.«» 3.» 3.» ».«»- 9.»:
».«» E» «.»» E» ».»» ».»» «.v» 8.»:
».«» e...- o.»» ».»» ».«» »..» «.»» .«-»:
».«» ».»» ».»» «. S ».»» E». .23.... .3.
«.«» 2:. »..» E» ».»» ».3 ».»» »Y»:
».«» »..». o...» a...» «.»» ».»». .6» «»-»:
».«» ».»» ».«» ».«» «.»» 3» c.» 8.»:
».»» - - - ».»» - »..»». a. .«8». \ 28.5.8». «83..
».»» »..»» »..» 3.» ».»» ».»» ».»» .»-»:
».»» «.»». ».«» ».«» «.»» ».»» ».«» 3.»:
«.»» ».»» »..» ».«» ».»» ».»v «.a- 2-»:
...»§6 82%»... 2.3.5.2 3.7.3.2 8...-..22 82.5.: 32%.: .8388.»

.EESE :03 5 vaoo» cc— 3 £203 vuom

 

. . 60...:th €03 78—.va map—yin? voom .N.< OBQH.

142

 

. . 60:52.8

 

143

 

»8« »8« 8.» »..« ».«». 88» .8» 8.2»... 8-»:
«.»« 8». 88 »8« «»»« »«»« 8.» 83.2». 8.»:
v«»« «8. .8» »8« »«»« .».« 88 3.2»... 8.»:
:»« .8. ».«» 88 »8« :»« »8« 2.9228982 »..»:
»8« »8« ».«» «»»« »8« »»»« .2» 8.2»... ».-»:
«»»« »8« ».»» «»»« 8». «.»« »»8 8.2»... ».-»:
»8« $.« »8» »».« .».« 8.» 88 .28025982 8-»:
.«»« 8». «.»» »8« «»»« »..» 8...» 8.2»... 8.»:
«»»« ».»« »8» »».« ««.« 8.» »«»» 8.2»... 8.»:
»8« 88 «8.. »8« :»« »».« .8» 83...». 8.»:
88 »».« ».»» 88 «»»« .«»« ».»» 8.2»: 8.»:
:»« «»»« .8» »»8 «8. 88 »8» 2.0.022382 8.»:
«8» »8« .8» 88 »».« 8.2 »«»» 8.2»... 8.»:
««8 .8. »«»» «»«» »8« »..» »«»» 8.2»... 8.»:
»«»» «8. »».« »«.» «»»« .»»« ».»» 8.2»... ».-»:
88 ».«« 88 8.» »8« .».» 88 8.2»... .«-»:
»»8 »8« .8» »..«» ».«« 88 «8» 23...... 8.»:
.8» :»« »8«» «.»« 2». »8« .«8 8.2»... 8-»:
.»8 .8. «8.. «.»» ».»« »..«» »..»» 8.2»: 8-»:
v. »» »8« 88 »8« »8« .8». 8. 2»... .88 .825. 8.8.
88 - - - «»»« »»v. »».v 8.2»... .2826
88 - - - - »8« 88 8.2»... .8...
.8» - - - - - .8» 8. 2»... .288». \ »«»»»... ». »..»».
2.226 8.22.2 8.22:. »8.-..22 8.222 82.5.: 88.22:. 2.». .6...» .8883.

.2283...» :08 5 7»: Eu;

 

.»8. .2.w »8. .88. .88. 2 29.5
2.52 .2» 282.22 .2222: 2 822» .88.? .8... .2» .« 8.228.. 8 28.2. 2.. .2» 3... 8 .o .73.»... .22..» .».< .2...»

. . 60.25.80

 

144

 

Namm 2b— mamm «.an G: oeuu 3cm v9. :94 3-3 _
vocN n2: 3 :V §N wwem 3cm NmNm co. Em...— msé— —
w SN 9%. NNNM nh—N. 80— 08m gun to. Em..— nvua —
_ 8N ocmm mmNm 3cm 2 S VNR .3— m v2 E 5.2 _
snow .. - .. anN - Nam v9. Em: scum-EU \ xoocfiov Xena—
hveN an _ mama Nam 8mm 33 van to. E»...— mmé— _
aveN eNn _ wowm 8am NNmN NNoN 32 v0.— E 3.2 _
NecN SE .»ch 2.2% 83 can nvum to. E oaé. _
cocN 3.3 32 Son own — cm-m wnnn v8 E 953882
econ 0mm: 3 on Sam mag 83 8mm v0.— E 2.3 _
meow mmm. mNOm nhwm awn cos-N :3 v8 2.3.4 wan-E _
nheN ¢oom when n _ am #3— chN wme v9. 2!..— 91m— _
gen 32 mNmm can 3mm 2 3 av; v9. 2»: 3.3 —
33 23 as.” 2.3 an cmvm Nmmm 3.. E 3-2 _
wok-N own _ c _ mm M: m n.3— aNnN moon v9. E 8.2—
wobm 3 S manm mnom ©3— amem sen to. E wmé— _
.»th 83 OS m wan as .NwN ”Nam 60.. Em..— 3d:
wmhm 38 can 33 5mm can. gem _a_o§o?:oz QWE _
Nos-N omen neon :3 2.2 23m wean v8 Em..— ow-E _
wok-N veg 2.5m «man .53 when chem v9. Em..— cN..a_ —
owa aim 3mm 3mm 85 $3 Nmnm v8 Em...— 3.3—
was 2mm 3mm vmom 0&— SEN mocm v9. EMS 3-3 —
Mann 3.— N 3mm .»ch NSN avg 83 3.. fun— nh-a_ _
mas-N as: new m Sam :3 2.: 32h Emogooéoz 8.3 _
2:330 33.; 83-522 33.—.22 82.-no.2 wag-:52 wag-:22 .530 voom .5_808<

 

...8=228 $2.84.. «22> .».< 02...»

. . .Bflflmafico

 

 

3% n3 3% man 82 5m 38 vein 3A.:
#3 32 :2. can - - - v8 a»: cicosgoifimuv 58x
SS in man $8 $2 $2 32 team: 8;:
82 new 38 $2 3.: $2 32 8:55 NE:
8% :2 N3. :8 $2 88 SR 85.5 3a:
a: on: ”a 88 :3 $2 33 _a_8§e$=cz :3:
E: ”8 gm «2 n can 82 an B. E :-a:
3% E $3 $3 ”a a: Rom 8:15 “3:
3% S: 88 :3 RR 82 8% 39558.52 93:
83. 22 3% S3 83 EN 38 83:5 9.2.
8: 8m 8% 23 «SN 2.: 3% gain 3;:
EA no em 83 3g 22 3: 83:»: NJ:
8: ”an $3. 8mm 38 RR 32 we “:5 93:
3mm 82 GS. 82 - - - we a»: uauxom \ xi: 3: 32.3.
82 can 2% 8% SR 22 h _ mm “.2 :5 8-2.
3.2 32 5m 89. - - - c2 2»: .902qu \ “.3: 86 833.
$2 a? :2 SS 28 32 man SE»: “M58
32 8: on: can as 82 2: 328.6382 33:
32 ma 8% an as $8 32 8.2»: 22:
$3 82 a; £8 82 22 a: v2 2m: ”2: _
EN 22 3a ”SN «3: 83 Nam 83:»: 8;:
:2 - - - Nae - s: c2 in 3%: Ba
83 a: 8mm 2% one as am we ”:5 «3:
an Sm 32 EN a: man 22” 85.5 as:
.2326 23352 §E22 ”8222 Ravi: 237.32 082.22 .28 33. .8332

 

 

..._§.§=8 9.2%: ”so; .2 03.:

145

motor: “—025 - a
gun—anon 0:. ma 3:88 - u
aqua—age ~26 vow-226 - :55 a

Anna—age a. .85 vomfieé 20$ :30; mam—Boone: 8 39.88 woman—.3 203 52:82:: - .

 

w_a_
cmag
mwau
a_oN
whom
nwoN
memN
wb_N
mNNN
aNNN
VVNN
anN
chN
thN

"N
~av_

new
N_m
own
Nun
mwv
ans
mmv
ovv
me_~
man
cNo
~on
cow
van—
vac

@—
been

when
aa—w
vcaN
«won
cw—n
mNMN
NmMN
moon
~th
v_mm
cva
ava
ammm
awwm
N_mm

h—
__5N

cth
mamm
ommm
~5NN
m¢-
_vn_
hmnN
«ohm
N—VN
cuou
thN
_amN
emam
mhh~
memm

@—
aa_N

~N¢N
www—
www—
95¢—
hang
an—N
show
thN
mca
ca-m
~o¢~
hmm_
«ww—
ONcN
e_¢N

@—
vnvN

enc—
awn—
h~_N
onc—
NceN
www—
wm-
-©N
man—
how—
v_m~
eNhN
ween

#—
ammm

”can
wean
cmmm
mwnm
wwwm
mNQN
mvau
___m
mmom
aawN
cmmm
when
mNoN

Gov :23? no 830530

:82
8353 3583.858 $33.
8.2»: 23:
.epsguseoz 3::
83:5 5::
83:5 :3:
303588.82 31¢:
83:»: 3.2.
83:»: mi:
:25 9;:
83:5 8d:
WEEOOICOZ Omua— —
8: 2»: 39856558 :88:
83:5 22:
83;»: 3-2—
83:5 $2.

 

o=95>O 33-; 32.532 33.534 82-532 3355: 33-—~22
393:0 nowo 5 A72...“

 

Eu;

.mouxvvoom uxzwmxxxx

...8==:=8 523$ 83> .n.< 9:3

146

. . 603.580

 

 

 

:8 8.: «.8 :8 :8 «.8 «.8 83:5 8-2.
».8 - - - - - ».8 83:»... g28.8. \ ««.«»... 22.8.
«.8 «.8 «.8 :8 ».8 «.«n «.8 83:»... .22.
«.8 n8 «.8 ».8 «.8 «.8 «.8 83:»... 8.2.
«8 m8 .8 :8 «.8 «8 «.8 83:»... 8-2.
«8 «.8 «.8 «.8 «.8 «.8 «.8 83:»... 2-2.
:8 E» «.8 «.8 «.8 «.8 «.8 83:5 ««-«:
«.8 «.8 :8 «.8 «.8 «.8 :8 83:»... 8-2.
«.8 «.8 «.8 ».8 «.8 «.8 ».8 83:»... 8-2.
«.8 - - - «.8 «.8 :8 83:»... .2856
«.8 «.8 :8 :8 «.8 «.8 «.8 83:»... 8-2.
.8 .8 :8 «.8 ».8 «:8 «.8 83:»... 8-2.
.8 «.8 «.8 :8 «.8 «.8 «.8 83:»... ««-«:
«.8 8.8 ».«« «.«m .8 :8 «.8 83:»... 8-2.
88 «.8 «.8 «.8 «.8 ».8 «.8 83:»... 8.2.
».8 «.8 ».«« «8 «.8 «.8 «.8 83:»... 8.2.
«.8 «.8 «..« 8.8 .8 «.8 ..«« 83:»... 8.2.
«.8 .8 ««.« «.8 :8 :8 «.8 83:»... ««-«:
«.8 «.8 «.8 «.8 «.8 .8 :8 83:»... 2-2.
.8 «.8 «.8 «.8 .8 «.8 :8 83:»... ««-«:
:8 - - - - «.8 «.8 83:»... .87..
:8 «.8 «.8 :8 2.. «.8 «.8 83:»... 8-2.
:8 «.8 ».E «..« :2- :8 .2 83:»... 8-2.
...§>o 82-3.52 «««.-::..2 82-8.2 32-8.2 82-5.2 82.8.: 25 80» .8832

888:...» 88 =. .788 8. .». .:».§ 8:»

.«««. :8 .»«2 .«««. .:««. a. 59.5 :82 :5
5805.2 .::».:«.2 c. :25... 88:9 3.8:: :5 .« 88.8.... 8 35.2. 9:. :5 «.5. 8 8 .88 .8. .». 3:».03 88 .:.< 3::

147

. . .voacvcoa

 

 

a.wm M.S. a. .v adv v.vv can adv v2 Ewm aYa:
a.wm N.vm n.2- Nfim v.vv mdn «awn EBEOQéOZ E.-a— _
adm w.vv Név _.am ~.vv Nam ficv v8 E n~.a_ _
can 03 v.mv Own v.vv Wan n. v v0.— ..39 Nma:
can «awe van _.vv _.am e. _v v._v v8 E 3.3—
can a.am wsv van n.av a.vv a.Nv v9. #39 vm-a_ _
man 03‘ man mfim a.vv m.nv Nev v2 .35 2a:
mam a.vv v.3 v.vv n.3- v.vm fimv v0.— 38 av.a_ _
mam a.a¢ a.vv v.vm flan 3v Ndv _a_o.—o:==oo.:o2 cva— _
mam v.~v adv a.vv N. _v o. .v adv v9. Em..— va-a:
man a. _ m a.vv v.~v fiwn a.vv adv v8 Ewfi 3a:
man Ndv a.vv adv n.wv a.vv a.vv _a_o..oEEoo.=oZ n _.a_ _
ham a.vv .Kv m.mn v.2. nén can v2 E»...— _m-a_ _
«man :4 a.mv a.vv m.vv 2h c.~v v0. Em: 91a:
mam a.av v.vv N.nv wan vfim n. .v to. Utah— Nva:
v.vv Qwv m.Nv van m.wv a.vv v.Nv fifibfificvécz moa— _
v.vv v. 3 Nam m.av w.~v «.an a-mv v8 E 3.2 _
adv _.vv NSv man man a.vv :.v v9. Em: mvua_ _
v.vv _.cm mdv ndv ”.mv _.¢v _.Nv v2 3.80 «Na:
v.vv QNm was v.vv m-hv v.vv a.vv v2 2.3..— _¢.a_ _
v.vv v.vw N.mv v.vv v.vv adv _.mv 60.. #80 30a— _
v.vv Ndm _.mv _. ~v a.vv a. .v N. _v v8 23.— .Na— _
v.vv a.vv Y:- o.am Ymv cav N-vv 8.. Emma vva— _
v.vv asv mswv v.vv v.vv can mév to. E Rwa— _
:=a..o>O aaa—$392 aaa 7:0va waa—éom—a aaa 7:32 vaa—éfi—Z vaa_-:o=2 0&0 voom $58303.

.5558.» :28 ... .788 8.». .:».2. :8»

...8..:....8 .83 _.8..»n. 35...... :8» .:.< 0.5.:

148

. . 6033.80

 

 

Sn ”.3 3n ”.2 E. - 3n Bias .258 “Egg 82.3.
new 34 2n _.8 23 in 8“ 8:18 ”3:
3m as” can ”.8 :6 at. «.8 vegan 8-2.
Sn 3:. «.8 3n 2% 3m 23 83:5 3-2.
”.8 9:. :3 9% as 9% e8 ‘22»: ”2:
Sn 3:. QR Sn «:3 in 9% use»: 2;:
.5. NS 9% 3m :8 «R «.2 320858.82 :3:
E 3. ed. EM «.8 2m 3m veg 23:
in 99. 3n 9% n3 2% 9% 39558.82 :1: _
«.3 a? 9% 3m «.8 E 3“ was»: 23:
in - - - as - in 83.5 .22: 3.
3“ Se a? 3n «.8 3m :3 nexus SJ:
0.: Se ”.8 3m - - - c2 .5: becazogsmuv swag
Em 3:. 5.3 as 3“ in :3 8:5: N3:
Sn «.8 96 2.“ 23 2n 9: 83:5 9%:
in man «.3 mm «.3 5.8 :6 83:3 3;:
3% Sn Se 2% «.8 Sc 9% ageeseoz e2:
:3 98 NS 2“ 9% 2n 2» 8:35 “2:
8m 8.. «8 Sn 5.8 3“ <8 8.2»: mg:
in S: 98 2e 9% So «.8 was»: $5:
33 - - - «.8 - ”.8 8.2»: bxdoiSsfiv :83.
Zn as. «8 m3 v.3 Sn 9% 8.2»: nix—mo
3n «8 5% 2n ”.8 new 25 8:35 5?:
3“ as. as new 9% 98 ”.8 veg 953282
2.55 $2-382 aarfiz 32.52: 58222 @8252 @8753 25 38 .5382

.2238 68 a p-83 2:3 2903 Bow

...B==€8 A88 _.8_ away»; Bow .«.< 2.3

149

mount? x35 - a

noun—anon 05 mo 85::— - u

anon—age 56 “.3826 - =Eo>O a

Engage“. =¢ 8.6 compass £303 v08 :20? €033 8 3288 woman—us 203 Eommmouu< - .

 

S 3 5. 3 G 3 8%? ho EsESo

3v «.8 2n «.3 :3 as :82

- we. fix «.2 «.8 - - c223 uAv—xgmulooczov 283.
Zn 0.8 v.8 2m «.3 2m 2m 3958.50.82 i:
can 3% 3m ”.2 can - - B. E»: booing—awe Sons.
9?. :3 MS 2.. - - - B. a»: .Aecxoh. \ V.5: 86 333.
on“ a? in 3a 98 a? «on o2 :3 :d:
3% ”.2 5a ”.8 as 2m ”.3 33555382 8-2.
2% 2:. 3m 98 «.8 3:. En 39585382 «2:
3m v.2. 9% 2m - - - v2 a»: 902qu \ is: 86 833.
new .3. 3% can ”.3 12 En we {.5 «3:
3m ma 0% 3m «.8 a? can 32055982 c2:
«.2 a? 3w 5% :8 3m «.2 we {5 v2:
«.2 0:. man 2% m. G 9% «.8 we £5 2;:
3n 3 «.8 9% 2m . 9: 0% we a»: oi:
Sn 2.. 0% can NR new m. s we £5 o3:
Sn nu. So «.3 NS 3; 0.8 we a»: 8;:

.235 32-; 82:22 @3222 28222 @8352 937.32 25 Bum .8532

.2235 :08 £ A738 2: .3 £303 Bum .
...B=€=8 €09. 72:3 3.3;; Bum .«.< 05.;

150

 

. . 60:52.8

 

 

.2» ».» 3 a.» m.» o.» a.» .~-»:
5.» e.» ..» a» e.» 3 n.» 3.»:
».» a.» 3 a.» 3 ».» m.» N»-»..
».» N.» 3 ».» 3 »..» 3 8.»:
».» ».» 3 ».» 3 m» ».N »..»:
».» ».N 3 3 3 3 ».» ».-»:
».» m.» 3 3 n.» N.» N.» 8.»:
».» ..» 3 3 3 - v.» 9.8.»: :88.
».» ».N 3 a.» 3 o.» 3 ».-»:
a.» 3 ».» N.» 3 o.» a.» 8.»:
a.» - - - a.» - - .6.-.» \ 5.8.82. 823.
a.» ».N a.» a.» 3 3 3 2-». .
a.» a.» 3 N.» 3 3 3 8-». .
a.» v.» 3 a.» - - - .333. . .3285 833.
3 a.» N.» 3 3 m.» 3 3-». .
3 - - - 3 - ».» .58». . 3». 838.
3 o.» 3 3 3 ».» a.» .8382
3 N.» ».» 3 3 3 3 8.»:
3 ».N 3 3 3 3 3 3-». .
3 3 3 3 - - - .320 . .328... 8:8.
o.» - - - - - o.» 9.5825223833. «~88.
v.» 3 m.» 3 o... - 5.» 9.35.3.
».» 3 5.» N... ».o G 3 8-»:
3:25 9.2-3.2 9.2-5.2 2.2.5.2 8.2.5.: 8.2-5.2 2.2.5.: .8388...

 

 

.83 8 33 88.. 88.5 5.52 v5 882.52 .camwnoaz 5 v8.8:—
.»oco_..a> x35 2.» ._ coca—:9..— mo 3:83 05 v5. 33 mu Eat 9.8; 3389... ac coca-aaa» .8 8.50m .m.< 03.:-

151

 

 

. . 62.5280

 

 

ad ad nd —.e dd _d ad 2a:
dd ad md _.e _.e ad ad dca:
dd md _d a.e ed ad dd ea-a:
dd ed ed ¢.e d.e ad _d aea:
dd dd _.e hd dd ad ad deua:
ed dd —.e ed dd ad e.e era. _
ed dd d.e od dd ¢.e cd daa:
ed ad pd md m.e ad dd ae.a:
ed ed dd ad d.e ed ed 593 _
ed ed _.e dd he ed ad ada— ~
ed ad _.e md c.e 5d dd ea-a:
ed dd fie c.e ed dd ed mea:
ed ad me ad ad dd nd mcua:
ed ad ad a.e ad dd _d dda _ _
ed ad d.e he .d ed md ea..a_ _
ed nd 9e ad ed 5d ed aaa:
ad ad ad a.e ad md ed :a:
ad ad ad 9e n.e ad ad dda:
ed _d m.e ad ad ad ad ada:
wd ad ad ad he ed ed aa-a_ _
wd dd fie dd nd nd ed Qua:
ed ad ad fin ad md hd ed-a:
5d 5d d.e a.e _.e ad md 8.3—
3:805 aaa—-UEQZ aaa—hum: aaa—£22 baa—£22 caaTEhZ oaa—AEE .g_303<

 

 

. . donates Ema; 35.8.5“ Sh mega .m.< 035.

152

 

 

. . 62.558
ad ad dd ad dd dd dd ae.a:
ad dd ad ad dd dd a._ dd-a:
ad ad d.e dd dd fid ed aaa:
ad dd ad ad fie dd dd dd.a:
ad fid ad ed ad ad dd da-a:
ad fid fid fid ad ad dd aaa:
fid ad dd dd fid dd ad de.a:
fid ad d.e ad ad fid ad da-a:
fid fid a.e d.d ed e.— dd ea-a:
fid ad fid ad dd ed dd cana:
fid a4 ad d.e dd dd ad aaa:
fid ad fid dd ad ad dd dTa:
fid dd ad dd ad dd ad ae.a:
dd ad d.e ad ed dd dd da-a:
dd dd ad dd dd ed ad aaa:
dd dd fie ed ad ad fid de.a:
dd ad dd ad ad ad ad _a.a:
dd ed e.e dd dd dd ad aa_-a:
dd ad ad ad dd ad ad ada:
dd fid a.e dd e.e fid ad ee.a:
dd dd ed ad ad fid fid ad..a:
dd dd dd ed ad dd dd ad..a:
dd ed ad dd ad dd ad dTa:
9:305 aaa—$302 aaa—hum: aaa—hum: aaa—Ava: aaa—-552 aaa—£32 .=o_$§<

 

 

:605538 Ems 3:88am.“ 8.. 830a .d.< oBaH

153

mound—g moo—5 - a

noun—anon 05 do 388.— - u

n§§>§ 56 ©0953 - =55;

Aagéfi aa 56 vows—95v 85.39% 3.. 8.88 :53 afivcoomov 8 3988 woman—.3 Eu? 828623 - .

 

 

2: 3. 33 S: 33 3: 8:69 3o 256880

3 3 3 3 3 3 :82
3 E 3 3 3 E 3 :3:
3 2 3 3 3 3 3 3:: _
3 3 S 3 3 3 3 3-»:
3 - 3 3 3 - 3 93 V28
3 3 3 3 3 3 3 8.3:
3 3 3 S 3 3 3 23:
3 3 3 3 3 3 3 3-3:
3 S 3 3 3 3 3 e3:
3 3 3 3 3 3 2 33:
3 3 3 3 3 3 3 :3:
3 3 3 3 3 3 3 33:
3 3 3 3 3 3 3 e3:
3 3 3 3 3 3 3 93:
3 3 3 3 3 3 3 :2: _
3 3 3 3 3 3 3 3»:

.=_E>o 3382 §7§2 «8222 575:2 3.15:2 23322 .5382

 

 

:63530 3.55 35.—comma .8.“ 380a .d.< 03:...

154

 

. . 60:52.8

 

 

3 3 3 3 3 3 3 :3:
3 3 3 3 3 3 3 m3:
3 3 3 3 3 3 3 .3:
3 3 3 3 3 3 3 . 33:
3 3 3 3 3 3 3 m3:
3 3 3 3 3 3 3 m3:
3 3 3 3 3 3 3 33:
3 3 3 3 3 3 3 e3:
3 3 3 3 3 3 3 m3:
3 3 3 3 3 - 3 ”3%: 3.83
3 3 3 3 3 3 3 a3:
3 3 3 3 3 3 3 33:
3 3 3 3 3 3 3 33:
3 3 3 3 3 3 3 a3:
3 - - - 3 - 3 .33: \ 03v 832
3 3 3 3 3 3 3 .5325:
3 3 3 3 3 3 3 :3:
3 - - - 3 - - 9633 \ 58.8% :38.
3 3 3 3 - - - agaxiaévoé 833
3 - - - - - 3 u§§8§=£8<§3a 3383.
3 3 3 3 - - - .35 \ 3286 838M
3 3 3 3 3 - 3 $333.
3 3 3 3 3 3 3 8.3:
2826 373—2 3222 3222 23222 082.52 @8232 .8303

 

 

.aaa_ 8 aaa 89a Sou—«D 582 v5 833m: @282 E 3.3—d
3:03 “.0vo v3 4 noun—anon ac 3:23 05 can add da 89a «anon 38095 ac mafia—mm ac oouaov 8a macaw .a.< 035.

155

. . .UOq—Cmun—OO

 

 

_d vd dd ad ad ad d._ odd:
_d ad 9v d.v ad od 0; «ed:
_d ed dd Nd ad ad :V 2.3—
Nd vd ad dd _d dd ad dad:
Nd dd ad vi dd dd dd NN.w:
Nd ad ad dd _.v dd dd avud:
Nd _d d6 Nd Nd dd _d Nd-w:
Nd _d dd e.e _d ad dd 2%:
Nd vd Nd ad Nd dd dd dcud:
Nd vd 9v dd d.v dd dd v1.»— _
dd Nd dd dd wd Nd dd wad:
dd ad dd dd Nd dd vd odd:
dd ad Nd dd 9v _d ad add:
dd od vé ad ad ad ad 8.x:
vd Nd e.e _d dd Nd od Nad:
vd od dd dd «6 ad _d wad:
vd ad ad ad _.v ad ad —a-w:
vd dd N6 Nd e.e vd vd aN-wZ
vd _.d dd ~.d ad a._ «d VNé:
dd dd a.v ad dd dd dd douw:
dd ad ad Nd 9v Nd vd dN-w:
dd dd e.e ad Né _d od acd:
dd ad _.v cd d6 dd dd 3.3—
a=80>0 aaa—.anZ aaa—-532 aaa—-532 aaa—£22 baa—$52 baa—£22 .=o_803<

 

 

...B=§=8 3:5 #53 a. 8M8 .3 «28m .3 03¢

156

 

 

. . 6025.80

 

 

ad dd Nd Nd dd ad d.— add:
ad dd dd ad dd _d dd NTd:
dd dd dd ad ad ad ad dN..d:
dd dd _d _.d ad ad ad _d..d:
dd dd Nd _d dd d.— d._ add:
dd _d ad ad dd dd dd ad:
ad dd dd _d dd d._ dd adud:
ad ad dd dd _d ad ad aa-d:
ad dd dd Nd Nd dd Nd aaTdZ
ad ad ad ad dd ad Nd dad:
ad Nd dd dd ad Nd Nd dd-d:
ad ad dd ad _d ad ad dd.d:
ad dd _d dd a._ Nd dd dad:
ad dd _d dd dd a4 dd dd-d:
ad _.N ad _.d dd d4 dd dad:
ad a._ ad dd dd d._ Nd dad:
_d dd _d dd dd dd dd ddd:
ad ad dd ad dd Nd Nd Nd.d:
.d dd dd dd ad dd dd add:
_d ad dd dd Nd dd dd add:
_.d dd Nd dd ad dd dd dd.d:
_d _.d dd dd ad ad dd Nd.d:
_d dd dd ad ad dd ad 3d:
a=Eo>O aaa—$392 aaa—£22 daa_-:o_2 aaa—-aaa.: daaTGEE daaTn22 .acmmmooo"

 

 

zdoacdcoo Ghana wry—5:3 do coho“. 3d macaw .d.< 055.

157

mount—3 «86 - a

8.383 23a 383.. - u

anon—noes .85 dodflga - =35;

Amazon—age =« .26 dump—26a 0323.“ 8d 8.58 :53 dam—Eng 8 dam—v.53“ dog 89.» 338034. - .

 

 

a.d_ N.a_ a.dN add d. _N _.NN noting do 2.66530

ad dd dd dd dd dd :32
ad d._ a._ ad dd d._ d._ dd:
_.N ad a._ dd ad d._ d._ —d-d:
_.N d._ dd dd Nd _.N d._ dd.d:
Nd d._ dd _d ad dd d._ dad:
dd dd dd d.~ dd d._ ad dad:
dd - dd dd dd - d._ uNd van—U
dd a._ dd dd dd dd d._ de:
dd Nd ad ad dd d._ d._ a_-d:
dd ad dd dd dd a.— d._ dd-d:
dd _d dd ad dd d._ d.— dd.d:
dd dd dd ad dd a._ a._ aN..d:
dd _.N ad Nd dd a._ dd :d:
dd ~d ad ad ad dd ad dd-d:
dd ad dd dd ad dd —.N add:
ad ad dd dd Nd dd a.— dd.d:

9:830 aaaE—an—Z aaaTsoE daa_-:o=2 aaa—£22 daaTGEE daa_-:o_2 .commmoood.

 

 

...B=§=8 Assassin do Es. «28m .3 as:

158

. . 62.538

 

 

3 3 3 3 - - - we 2m: uneven \ 3336 :33.
3 3 3 3 3 3 3 BE»: 33:
3 3 3 3 3 3 3 BE»: 33:
3 3 3 3 3 3 3 was»: $3:
3 - 3 3 - - - 8. 2»: 93855338 :33—
3 3 3 3 3 3 3 3235 ml:
3 3 3 3 3 3 3 8:35 3.2.
3 3 3 3 3 3 3 3.2»: :3:
3 3 3 3 3 - - ‘22»: 69856338 «33.
3 3 3 3 3 3 3 322»: 5.3
3 3 3 3 3 3 3 8:25 2-2.
3 3 3 3 3 3 3 8:35 “3:
3 3 3 3 3 3 3 was»: 33:
3 3 3 3 3 3 3 3230 08.832
3 - 3 3 3 3 3 8:15 3:
3 3 3 3 3 3 3 8:4»: 832
3 3 3 3 3 3 3 8:5: 33:
3 3 3 3 3 3 3 322»: N3:
3 - 3 3 3 3 3 ‘22»: 83:
3 3 3 3 3 3 3 8:3— 33:
3 3 3 3 3 3 3 was»: ”3:
3 3 - - 3 - 3 33:5 .2323.
3 3 3 3 3 3 3 83:5 33:
2.55 82-32 82.5.: 3222 3222 37:52 3753 25 Bow .8383.

 

 

.aa_ 3 daa_ 89d 88—5 5.82 d5 88352
.3222 E dean—a .8305; x85 33 .N sang—aaa do 3:83— 05 d5 35— da 89d 33d @3809:— do 8:38am.» .5.“ macaw .a.< 035.

159

. . 6025.50

 

160

 

3 3 3 3 3 3 3 veg n3:
3 3 3 3 3 3 3 75:5 33:
3 - - - - 3 3 BE»: e83
3 3 3 3 3 3 3 33:28.52 83:
3 3. 3 3 3 .3 3 8:25 «3:
3 3 3 3 3 3 3 32»: 33:
3 E E 3 3 3 3 _a_au§=$=oz «3:
3 3 3 3 3 3 3 83:»: c3:
3 3 3 3 3 3 3 8:35 “.3:
3 3 3 3 3 3 3 83:8 03:
3 3 3 3 3 3 3 Bio 23:
3 3 3 3 3 3 3 83:5 3.3:
3 3 3 3 3 - - n22»: ufifimUZSEQ 333
3 3 3 3 3 3 3 c2233 33:
3 3 3 3 3 3 3 83:»: a}:
3 3 3 3 3 3 3 @235 SA:
3 3 3 3 3 3 3 8.2»: $3:
3 3 3 3 3 3 3 we 2m: 23:
3 3 3 3 3 3 3 8:55 33:
3 3 3 3 3 3 3 83:»: and:
3 3 3 3 3 3 3 8.2»: 338
3 3 3 3 3 3 3 95:5 Sa:
3 - - - 3 - 3 “.223: 9958;85:an «33
.555 82.; 82-5% 82-55 38222 @8753 23222 25‘ 88 .8302

 

 

.. 63.5.80 9.33 85.825“ .8.“ 8.80m .N..< 05a...

 

161

 

. . floss—«Eco
3 3 3 3 3 3 3 83;»: 83:
3 3 3 3 3 3 3 33:5 :3:
3 3 3 3 3 3 3 83:8 33:
3 S 3 3 3 3 3 33:5 33:
3 3 3 3 3 M: 3 B36: 33:
3 3 3 3 3 3 3 33382 33:
3 3 3 3 3 3 3 83:5 33:
3 3 3 3 - - - .3235: 35833285 833
3 3 3 3 3 3 3 83:5 33:
3 3 3 3 3 3 3 B: 3:5 N3:
3 3 S 3 3 3 3 93:8 Rd:
3 3 3 3 3 3 3 BE»: 3:
3 3 3 3 3 3 3 v2.5: :3:
3 3 3 3 3 3 3 33:8 33:
3 3 3 3 3 3 3 233.5382 :3:
3 3 3 3 3 3 3 83:5 33:
3 3 3 3 3 3 3 83.3: 3.3:
3 3 3 3 3 3 3 33:5 33:
3 3 3 3 3 3 3 B: a»: 2-3
3 3 3 3 3 3 3 8.3»: v3:
3 3 3 3 3 3 3 33:5 33:
3 3 3 S 3 3 3 we 2»: ml:
3 3 3 3 3 3 3 3358382 SA:
.2890 33352 3222 38222 3222 37:52 3222 25 Bow .8_§8<

 

 

. . 626.550 3&5 35.—«25¢ .5.“ «0.80m. .n.< 03$.

”0:35, “—005 - _.

8388 22.. $85 - .

323:0 ~26 “009205 - 385..

3333 E" ~26 vow—:95 003093 .8. «088 :83... 3083 9 «£6308 @0335 0003 333000< - .

 

 

3_ m: 33 n: 3.: 3. 885:288580

3 3 3 3 3 3 822
S E 2 3 3 2 3 3225382 3;:
S 3.2 2 3 S 2 3 39588982 33:
3 3 E 3 S 3 3 3288382 83:
3 3 3 3 3 3 3 36882382 33:
3 - - - - 3 3 3 2»: E33 \ 3330 333.
3 3 3 3 3 3 3 32»: 23:
3 3 3 3 3 3 3 338382 8.2.
3 - 3 3 3 3 3 3288382 33:
3 3 3 3 3 3 8. a»: 9883 0.8560 .33
3 3 3 3 .3 3 3 3.8: 83:
3 3 3 3 3 3 3 3.8: 33:
3 3 3 3 3 2 3 8. .8: :32
3 3 3 3 3 3 3 3.8: 83:
3 3 3 3 3 3 3 3.8: :8:
3 3 3 3 3 3 3 c2089 33:
3 3 3 S 3 3 3 8:8: :3:
3 - - - 3 - 3 8:83 38:20

9:830 887322 82-83 ”378.2 §78§ 887822 @8732 25. 8% .3883

 

 

zdoacucoo Emit 005.33% 8.2 8.80m .h.< 038.

162

. . 62.5250

 

 

163

3 3 3 3 3 3 3 83mm: «3:
3 3 3 3 3 3 3 8.2»: 3;:
3 3 3 3 3 .3 3 wei: ml:
3 3 3 3 3 3 3 gain: 23:
3 3 3 3 3 3 3 3:35 “3:
3 - 3 3 - - - B. 2»: 328563638 .33
3 3 3 3 3 3 3 ‘22»: 9%:
3 3 3 3 - - - B. a»: usage» \ x3285 3&3
3 .3 3 3 3 3 3 83:5 :3:
3 3 3 3 3 3 3 was»: :5:
3 3 3 3 3 3 3 vegan 35382
3 3 3 3 3 3 3 3.2»: 8.3:
3 3 3 3 3 3 3 c2 a»: mi: _
3 3 3 3 3 3 3 ‘22»: e3:
3 - 3 3 3 3 3 83:5 93:
3 3 3 3 3 3 3 8:55 33:
3 3 3 3 3 - - we a»: 338553.538 233.
3 3 3 3 3 3 3 BE»: 33:
3 - 3 3 3 3 3 92.53 83:
3 3 - - 3 - 3 8:35 2353
3 3 3 3 3 3 3 322m: «3:
3 3 3 3 3 3 3 8.25 ”3:
3 3 3 3 3 3 3 8.4.5 33:
.2826 82.3 38222 ”$222 5222 375.2 3232 25. Bow .8503.

 

 

.82 2 32 52. 3.5 552 as. 38852 .3222
5 @853 .8303 38:0 98 .N gun—sac.— ..c 3:83 05 can add Q. EB”. 389 388.5 mo wart—gm uc 8&3 .8.“ 8.50m .w.< 033.

. . .Ugafimucco

 

164

 

3 2 w.” 5 2 tn 3 BEE: «3:
an 3 an «a 2 ca 2 3:5 «3:
2 Z 3 3 S. 2 3 BE «3:
an E 5 an 2, 3. .2 3955382 :3:
3 0.n 3. on «a Z 3 39558.82 Sé:
on an an 3 3. 3 2 vein e2:
3 2 3 2 3 ma 3 328550.52 mg:
on «a 2 3 3 3 a." 3:35 2;:
2 an on 3” S 3 3 83:5 Em:
2.. ea 3 2 9m 3 an 83:»: 2.2.
2 ca 3 2 an 3 3 vegan 22:
3 3 an 3 2 «a 2 axis a3:
3“ 2 an S 3 3 3 95:5 2.2.
3” mm 2. an an an 3 v2 a»: £3:
«a 3 mm 3 9n 2 3 8.2»: 3538
i 2. 3 2 3 3 3 neg»: 9;:
X 3 3 3 9n 3 «N 3922382 :3:
tn 3 2 3 2 9m 2 8:5: 3;:
3 2 3 3. 3 3 2” gain: a2:
3 2 3V 3 3 2 in gain: a}:
0.n - - - 3 - 3 BE»: uqumuZSssov X83.
0.n «a S 3 2. ca 2 8315 SJ:
0.n 3. on «a i - - was»: ufifimuloosfiv 83.3.
2890 82.; gavfiz ”8222 23755 2325: e822: 25 35 22883

 

 

...B=§§ Q59 95% 13 2.qu .3 850m .3 93.;

. . 63.5.80

 

 

165

3 - - - - 3 3 33;»: .83
e3 3 3 3 3 3 S 3&5 a3:
3 3 3 S 3 3 3 2.5.»: 2-2.
3 3 e3 3 3 3 2 8:18 :3:
3 3 2 3 3 3 3 BE»: 8;:
3 3 3 3 3 3 3 83:5 3.2.
3 3 3 3 3 3 3 Exam «3:
3 M: 3 3 3. 3 3 we £5 8-2.
3 3 3 3 3 3 ca vein ”3:
mm 3 3 3 3 3 3 v8 a»: 8-2 _
3 3. 3 3 3 3 3 vegan 2d:
3 3 3 3 3 3 3 “.2 vE5 m2:
3 3 3 3 3 3 3 3:25 Rd:
3 3 3 3 3 3 «N we a»: 8-2 _
3 3 3 3 2 3 3 8. id 33:
3 3 3 3 3 3 3 we a»: 21: _
3 3 3 3 3 S 3. 32358.82 :3:
3 3 3 3 - - - we a»: 995mm \ 83286 833—
3 2 3 3 3 3 3 3...»: 2.2.
3 2 3 3 3 3 3 Exam «2:
3 3. 3 S 3 3 3 83%: a}:
3 3 3 3 3 3 3 Exam 5a:
3 3 3. S. 3 2 E 8:25 3;:
9:36 aaiflz 82-53 ”3222 28222 32.55 23222 25 30m 22383.

 

“58:83.5 £043 5 Chamv Exam no get how 880m

 

. . fies—Esau Gummy wars—mm mo 02mg 8.. 8.50m .w.< 035.

mounts, ~86 .. a

cog—=92— 95 .3 fig.— - o

anus—age... 83 cows? - =§>O..

32838 .7 56 vowEogv 3582—9 3.. 838 :35 395806 8 95:83 3mg 82: Eomuao8< - .

 

 

a: 3: N: a: 2a SN =88§§§§80

2 3“ mm 2 3 3 :82
B 2 2 S 3 2 2 32258.82 2;:
3 E 2 2 Z 2 3 3955382 m3:
3 - - - - ta 3 B. 2»: L :33— \ $38: 23%
2 f 3 on 2 2 E 32053qu 31a:
2 S E 3. 2 2 3 3555382 93:
2 3 E 3 2 S S 3:55 :3:
2 _.N 2 an 3 E S 395588.82 8.2.
2 - on 3 ma 3 2 39555382 9-2 _
Eu 2 3 Z 2 2 _.~ 8.2»: 8;:
E 3 3 Z 3 - M: wei: uSSN 385:8 88$.
3 - - - tn - 3 BE»: .2856
2 Z 2 3 Z 3 3 we a»: a2:
3 3 ma 3 2 3 3 was»: ”2:
3 S 3 3 3 .2 E 3%: 3-2.
3 «a 2 3 3 2 S 8:18 vi:
3 3. 3 S 2 3 2 was»: 34.:
a.” 3 3 Z 3 2 2 3.2»: ”2:

..=§>o §_%oz aarfiz ”3222 5222 32.85 @8322 w5 Bow 22382

 

“58:83:”. :08 E .Smm £3: mo 353 8a goom

 

 

. . 603558 AhqmmmSE—nm .«o oomeu é 3.60m .w.< 035.

166

Table A.9. ANOVA for yield (kg-ha") of 2 replications of Population 1, planted in six
environments in Michigan, Minnesota and North Dakota from 1996 to 1999, used to
estimate heritability.

 

Source Df Expected Mean Squares Mean Squares
Environment, G v -— 1 41652120
Rep, R (r- 1) v 2971499
Genotype. G g —1 62 + 1.99 029;? 11.95 62, 1050286
GxE (g- l)(v- 1) 02+ 1.99 62,. 397258
Error v(g— 1)(r— 1) 62 186815

 

Table A.10. ANOVA for seed weight (g-100") of 2 replications of Population I, planted
in six environments in Michigan, Minnesota and North Dakota from 1996 to 1999, used
to estimate heritability.

 

 

Source Df Expected Mean Squares Mean Squares
Environment, V v — 1 6650.8
Rep, R (r — 1) v 74.5
Genotype, G g —1 <52 + 1.99 oz,.,+ 11.95 oz, 83.6
GxE (g—le—l) Uzi-1.9902,, 35.3
Error v(g— 1)(r— 1) 62 10.3

 

Table A.11. ANOVA for yield (kg-ha") of 2 replications of Population 2, planted in six
environments in Michigan, Minnesota and North Dakota from 1996 to 1999, used to
estimate heritability.

 

 

Source Df Expected Mean Squares Mean Squares
Environment, V v — 1 74795711
Rep, R (r - 1) v 2806308
Genotype. G g —1 0’2 + 2.00 62,.+ 11.98 0'2. 1085626
GxE (g- l)(v- 1) 62+2.00 62,, 407105

Error v(g— 1)(r— 1) 0’2 183115

 

167

Table A.12. ANOVA for seed weight (g- l 00") of 2 replications of Population 2, planted
in six environments in Michigan, Minnesota and North Dakota from 1996 to 1999, used
to estimate heritability.

 

Source Df Expected Mean Squares Mean Squares
Environment, V v — 1 5482.2
Rep,R (r-1)v 81.9
Genotype, G g —1 62 + 2.00 029+ 11.98 6’. 111.7
GxE (g- l)(v—1) 02+2.00 62,. 34.1
Error v(g—l)(r-l) 62 9.6

 

Table A.13. ANOVA for scores on appearance (APP) of processed beans of 2
replications of Population 1, planted in six environments in Michigan, Minnesota and
North Dakota fi'om 1996 to 1999, used to estimate heritability.

 

 

Source Df Expected Mean Squares Mean Squares
Environment, v v — 1 22.4907248
Rep, R (r - 1) v ' 1.9878034
Genotype. G g -1 o2 + 1.96 o’..+ 11.75 6’, 3.4476766
GxE (g— l)(v— 1) 61+ 1.97 6”,, 0.5986469
Error v(g- l)(r— 1) 0'2 0.4029635

 

Table A.14. ANOVA for scores on degree of splitting (SPLT) of processed beans of 2
replications of Population 1, planted in six environments in Michigan, Minnesota and
North Dakota from 1996 to 1999, used to estimate heritability.

 

 

Source Df Expected Mean Squares Mean Squares
Environment, V v — 1 29.8749153
Rep, R (r — 1) v 3.8871539
Genotype. G g -1 o2 + 1.96 629+ 11.75 oz, 3.8115691
G x E (g - l)(v — 1) 0'2 + 1.97 62,, 0.6354538

Error v (g — l)(r — 1) 0’2 0.4188847

 

168

Table A.15. ANOVA for scores on appearance (APP) of processed beans of 2
replications of Population 2, planted in six environments in Michigan, Minnesota and
North Dakota from 1996 to 1999, used to estimate heritability.

 

 

Source Df Expected Mean Squares Mean Squares
Environment, V v — 1 5.7590902
Rep, R (r— 1) v 2.1265000
Genotype, G g —1 0'2 + 1.96 029,“? 11.70 oz, 4.9363941
G x E (g - 1)(v — 1) o2 + 1.98 62,, 0.7669883
Error v (g — l)(r — 1) 0'2 0.3543357

 

Table A.16. ANOVA for scores on degree of splitting (SPLT) of processed beans of 2
replications of Population 2, planted in six year-location combinations in Michigan,
Minnesota and North Dakota from 1996 to 1999, used to estimate heritability.

 

 

Source Df Expected Mean Squares Mean Squares
Environment, V v — 1 16.8613975
Rep, R (r — 1) v 1.6999876
Genotype, G - g —1 62 + 1.96 613,)r 11.70 (,2, 4.9774993
G x E (g — l)(v — 1) o2 + 1.98 62,. 0.7503643

Error v (g — l)(r — 1) 0'2 0.3404407

 

169

LITERATURE CITED

170

LITERATURE CITED

ADAM-BLONDON, A.F., M. Sevignac, and M. Dron. 1994. A genetic map of common
bean to localize specific resistance genes against anthracnose. Genome. 37: 915-
924.

ADAMS, M.W. 1967. Basis of yield component compensation in crop plants with
special reference to the field bean, Phaseolus vulgaris. Crop Sci. 7(5): 505-510.

ADAMS, M.W. and CL. Bedford. 1975. Breeding food legumes for improved
processing and consumer acceptance properties. In: M. Milner (ed.). Nutritional
Improvement of Food Legumes by Breeding. John Wiley and Sons, N.Y. pp.
299-304.

ALLARD, R.W. and AD. Bradshaw. 1964. Implications of genotype-environmental
interactions in applied plant breeding. Crop Sci. 4: 503-508.

AL-MUKHTAR, RA. and DP. Coyne. 1981. Inheritance and association of flower,
ovule, seed, pod, and maturity characters in dry edible beans (Phaseolus vulgaris
L.). J. Amer. Soc. Hort. Sci. 106 (6): 713-719.

BAI, Y., T.E. Michaels and KP Pauls. 1996. Identification of RAPD markers linked to
bacterial blight resistance genes in Phaseolus vulgaris L. Ann. Rep. Bean
Improvement Cooperative. 39: 164-165.

BEAVER, J .S., C.V. Paniagua, D.P. Coyne, and GP. Freytag. 1985. Yield stability of
dry bean genotypes in the Dominican Republic. Crop Sci. 25(6): 923-926.

BEEBE, S.E., I. Ochoa, P. Skroch, J. Nienhuis and J. Tivang. 1995. Genetic diversity
among common bean breeding lines developed for Central America. Crop Sci.
35: 1178-1183.

BOROJEVIC, S. and W.A. Williams. 1982. Genotype x environment interactions for
leaf area parameters and yield components and their effects on wheat yields. Crop
Sci. 22: 1020- 1025.

BOTSTEIN, D., R.L. White, M. Skolnick and R.W. Davis. Construction of a genetic
linkage map in man using restriction fragment length polymorphisms. Am. J.
Hum. Genet. 32: 314-. as cited by MD. Burow and T.K. Blake. 1998.
Molecular tools for the study of complex traits. In: Molecular dissection of
complex traits. A.H. Paterson (ed.). New York: CRC Press. pp. 13-29.

171

BUROW, MD. and T.K. Blake. 1998. Molecular tools for the study of complex traits.
In: Molecular dissection of complex traits. A.H. Paterson (ed.). New York:
CRC Press. pp. 13—29.

CATELL, R.B. 1965a. Factor analysis: An introduction to essentials. I. The purpose
and underlying models. Biometrics. 21: 190-215. as cited by G.L. Hosfield, A.
Ghaderi and M.A. Uebersax. 1984. A factor analysis of yield and sensory and
physico-chemical data from tests used to measure culinary quality in dry edible
beans. Can. J. Plant Sci. 64: 285-293.

CATELL, R.B. 1965b. Factor analysis: An introduction to essentials. II. The role of
factor analysis in research. Biometrics. 21: 405-435. as cited by G.L. Hosfield,
A. Ghaderi and M.A. Uebersax. 1984. A factor analysis of yield and sensory and
physico-chemical data from tests used to measure culinary quality in dry edible
beans. Can. J. Plant Sci. 64: 285-293.

CHAGUE, v., J.C. Mercier, M. Guénard, A. dc Courcel and F. Vedel. 1997.
Identification of RAPD markers linked to a locus involved in quantitative
resistance to TYLCV in tomato by bulked segregant analysis. Theor. Appl. Genet.
95: 671 -677.

CHALMERS, K.J., U.M. Barua, C.A. Hackett, W.T.B. Thomas, R. Waugh and W.
Powell. 1993. Identification of RAPD markers linked to genetic factors

controlling the milling energy requirement of barley. Theor. Appl. Genet. 87(3):
314-420.

CHENG, F.S., N.F. Weeden and S.I(. Brown. 1996. Identification of co-dominant
RAPD markers tightly linked to fruit skin color in apple. Theor. Appl. Genet. 93:
222-227.

COCHRAN, w.G. and GM. Cox. 1968 Experimental Designs. 2'“- Edition. New
York: John Wiley & Sons, Inc. 611 p.

COPELAND, L.O., and M.H. Erdmann. 1977. Montcalm and Mecosta: Halo blight
tolerant kidney bean varieties for Michigan. MSU Cooperative Extension Service
Bull. 95 7 No. 81.

COYNE, DP. 1968. Correlation, heritability and selection of yield components in field
beans, Phaseolus vulgaris L. Proc. Am. Soc. Hortic. Sci. 93: 388-396.

DARVASI, A. and M. Soller. 1994. Selective DNA pooling for determination of linkage

between a molecular marker and a quantitative trait locus. Genetics. 138: 1365-
1373.

172

DESI-IPANDE, S.S., S.K. Sathe and D.K. Salunkhe. 1984. Dry beans of Phaseolus: A
review. Part 3. Processing. CRC Crit. Rev. Food Science and Nutrition. 21 (2):
137-195.

DOGANI.AR, S., S.D. Tanksley and M.A. Mutschler. 2000. Identification and
molecular mapping of loci controlling fruit ripening time in tomato. Theor. Appl.
Genet. 100(2): 249-255.

DRAKE, SR. and BK. Kinman. 1984. Canned dry bean quality as influenced by high
temperature short time (HTST) steam blanching. J. Food Science. 49(5): 1318-
1320.

DUDLEY, J .V. 1993. Molecular markers in plant improvement: manipulation of genes
affecting quantitative traits. Crop Sci. 33: 660-668.

EATHINGTON, S.R., J .W. Dudley and GK. Rufener II. 1997. Usefulness of marker-
QTL associations in early generation selection. Crop Sci. 37: 1686-1693.

EDWARDS, K., C. Johnstone, and C. Thompson. 1991. A simple and rapid method for

the preparation of plant genomic DNA for PCR analysis. Nuc. Acids Res.
19(6): 1349. ‘

ELIA, F.M., G.L. Hosfield, J .D. Kelly and M.A. Uebersax. 1997. Genetic analysis and
interrelationships between traits for cooking time, water absorption, and protein
and tannin content of Andean dry beans. J. Amer. Soc. Hort. Sci. 122 (4): 512-
518.

FEHR, W.R. 1987. Principles of Cultivar Development. Vol. 1: Theory and Technique.
New York: Macmillan Publishing Company. 536 p.

FORNEY, A.K., D.E. Halseth and W.C. Kelly. 1990. Quality of canned “Ruddy" kidney
beans as influenced by planting date, harvest time and length of storage before
canning. J. Amer. Soc. Hort. Sci. 115(6): 1051-1054.

FRANCO, M.C., S.T.A. Cassini, F.C. Montrazzi, S.M. Tsai and C. Vieira. 1998. RAPD
analysis of common bean (Phaseon vulgaris L.) cultivars and evaluation of
common bacterial blight (CBB) and wild fire (WF) resistance. Ann. Rep. Bean
Improvement Cooperative. 41: 143-144.

FREEMAN, G.H. 1973. Statistical methods for the analysis of genotype-environment
interactions. Heredity. 31: 339—354.

FREYRE, R., P.W. Skroch, V. Geffi'oy, A.-F. Adam-Blondon, A. Shirmohamadali, W.C.

Johnson, V. Llaca, R.O. Nodari, P.A. Pereira, S.-M. Tsai, J. Tohme, M. Dron. J.
Nienhuis, C.E. Vallejos and P. Gepts. 1998. Towards an integrated linkage map

173

of common bean. 4. Development of a core linkage map and alignment of RFLP
maps. Ihero. Appl. Genet. 97:847-856.

GABRIEL, O. 1971. Locating enzymes on gels. Methods in Enzymology. 22: 578- . as
cited by MD. Burow and T.K. Blake. Molecular tools for the study of complex
traits. In: Molecular dissection of complex traits. A.H. Paterson (ed.). New
York: CRC Press. pp. 13-29.

GHADERI, A., G. Varner and W. Adams. 1980. Adaptability of dry bean classes and
varieties to Michigan. Mich. Dry Bean Digest. 5(3): 12-14.

GHADERI, A., M.W. Adams and A.W. Saettler. 1982. Environmental response patterns
in commercial classes of common bean (Phaseolus vulgaris L.). Theor. Appl.
Genet. 63: 17-22.

GHADERI, A., G.L. Hosfield, M.S. Adams and M.A. Uebersax. 1984. Variability in
culinary quality, component interrelationships, and breeding implications in navy
and pinto beans. J. Amer. Soc. Hort. Sci. 109:85-90.

GONZALES, A.R., K.M. Edwards, and DB. Marx. 1982. Storage and processing
quality of beans (Phaseolus vulgaris L.) harvested at the semi-dry stage. J. Amer.
Soc. Hort. Sci. 107(1): 82-86.

GOTTLIEB, L.D. 1982. Conservation and duplication of isozyrnes in plants. Science.
216: 373-380.

HALEY, S.D., P.N. Miklas, J .R. Stavely, J. Byrum and J .D. Kelly. 1993. Identification
of RAPD markers linked to a major rust resistance gene block in common bean.
Iheor. Appl. Genet. 86: 505-512.

HAWTIN, G.C., K.O. Rachie and J .M. Green. 1977. Breeding strategy for the
nutritional improvement of pulses. In: Nutritional Standards and Methods of
Evaluation for Food Legume Breeders. J .H. Hulse, K.O. Rachie and L.W.
Billingsley (eds.) Ottawa: IDRC. 100 p.

HEIL, J.R., M.J. McCarthy and M. Ozilgen. 1992. Parameters for predicting canning
quality ofdry kidney beans. J. Sci. Food Agric. 60(4): 519-523.

HELENTJARIS, T., T .M. Slocum, S. Wright, A. Schaefer and J. Nienhuis. 1986.
Construction of genetic linkage maps in maize and tomato using restriction
fiagment length polymorphisms. Theor. Appl. Genet. 72: 761-769.

HELLER, R., J. Schondelmaier, G. Steinrllcken and C. Jung. 1996. Genetic localization

of four genes for nematode (Heterodera schachtii Schm.) resistance in sugar beet
(Beta vulgaris L.). Iheor. Appl. Genet. 92(8): 991-997.

174

HILL, J. 1975. Genotype-environment interactions — a challenge for plant breeding. J.
Agric. Sci, Camb. 85: 477-493.

HOFF, J .E. and PE. Nelson. 1965. An investigation of accelerated water-uptake in dry
pea beans. Res. Prog. Rept. 211. Agric. Expt. Station. Purdue University, West
Lafayette, Indiana.

HOSFIELD, G.L. 1991. Genetic control of production and food quality factors in dry
bean. Food Technology. 45: 98-103.

HOSFIELD, G.L. and M.A. Uebersax. 1980. Variability in physico-chemical properties
and nutritional components of tropical and domestic dry bean germplasm. J.
Amer. Soc. Hort. Sci. 105: 246-252.

HOSFIELD, G.L. and M.A. Uebersax. 1990. Culinary quality in dry bean: can it be
improved?. Ann. Rep. Bean Improvement Cooperative. 33: 17-18.

HOSFIELD, G.L., A. Ghaderi and M.A. Uebersax. 1984a. A factor analysis of yield and
sensory and physico-chemical data from tests used to measure culinary quality in
dry edible beans. Can. J. Plant Sci. 64: 285-293.

HOSFIELD, G.L., M.A. Uebersax and T.G. Isleib. 1984b. Seasonal and genotypic
effects on yield and physico-chemical seed characteristics related to food quality
in dry, edible beans. J. Am. Soc. Hort. Sci. 109:182-189.

HOSFIELD, G.L., J.D. Kelly, M.J. Silbemagel, J .R. Stavely, M.W. Adams, M.A.
Uebersax and G.V. Varner. 1995. Eight small-red dry bean germplasm lines with
upright architecture, narrow profile, and short vine growth habit. HortScience.
30(7): 1479-1482.

JOHNSON, E., P.N. Miklas and J .R. Stavely. 1994. The potential of coupling and
repulsion phase RAPD markers for indirect selection of rust resistant progeny in
common bean. Ann. Rep Bean Improvement Cooperative. 37: 81-82.

JUNEK, J .J ., W.A. Sistrunk and MB. Neely. 1980. Influence of processing
methodology on quality attributes of canned dry beans. J. Food Sci. 45:821-824.
as cited by W. Lu and KC. Chang. 1996. Correlations between chemical
composition and canning quality attributes of navy bean (Phaseolus vulgan's L.).
Cereal Chemistry. 73 (6): 785-787.

JUNG, G., D.P. Coyne, P.W. Skroch, J. Nienhuis, E. Amaud-Santana, J. Bokosi, H.M.
Ariyarathne, J .R. Steadman, J .S. Beaver and SM. Kaeppler. 1996. Molecular

markers associated with plant architecture and resistance to common blight, web
blight, and rust in common beans. J. Amer. Soc. Hort. Sci. 121(5): 794-803.

175

KAYS, S.J., J .W. Williams and DR. Davis. 1980. Harvest of dry beans in the pre-dry
stage of development: effect on yield and processed quality product. J. Amer. Soc.
Hort. Sci. 105 (1): 15-17. '

KELLY, J .D. and RN. Miklas. 1998. The role of RAPD markers in breeding for disease
resistance in common bean. Molecular Breeding. 4: 1-1 1.

KELLY, J .D. and RN. Miklas. 1999. Marker-assisted selection. In: Common bean
improvement in the twenty-first century. vol. 7: Developments in plant breeding.
S.P. Singh (ed.). London: Kluwer Academic Publishers. pp. 93-123.

KELLY, J .D., J .M. Kolkman and K Schneider. 1998. Breeding for yield in dry bean
(Phaseolus vulgaris L.). Euphytica. 102: 343-356.

KELLY, J .D., K.A. Scheiner and J .M. Kolkman. 1999. Breeding to improve yield. In:
Common bean improvement in the twenty-first century. S.P. Singh (Ed). Kluwer
Academic Publishers, Netherlands. pp. 185-222.

KENNARD, W.C., K. Poetter, A. Dijkhuizen, V. Meglic, J. Staub and M. Harvey. 1994.
Linkages among RFLP, RAPD, isozyme, disease resistance, and morphological
markers in narrow and wide crosses of cucumber. Theor. Appl. Genet. 89:42-48.
as cited by J .E. Staub and F .C. Serquen. 1995. Genetic markers, map
construction and their application in plant breeding. HortScience. 31 (5): 729-740.

KIM, J. 1975. Factor analysis. In SPSS: Statistical package for the social sciences. New
York: McGraw-Hill, Inc. pp. 468-514. as cited by G.L. Hosfield, A. Ghaderi and
M.A. Uebersax. 1984. A factor analysis of yield and sensory and physico-
chemical data fi'om tests used to measure culinary quality in dry edible beans.
Can. J. Plant Sci. 64: 285-293.

KNAPP, S.J., W.W. Stroup, and WM. Ross. Exact confidence intervals for heritability
on a progeny mean basis. Crop Sci. 25: 192-194.

KNAPP, SJ. 1998. Marker-assisted selection as a strategy for increasing the probability
of selecting superior genotypes. Crop Sci. 38: 1164-1174.

KON, S. and D. W. Sanshuck. 1981. Phytate content and its effect on cooking quality of
beans. J. Food Processing and Preservation. 5(3): 169-178.

LANDER, ES. and D. Botstein. 1989. Mapping Mendelian factors underlying
quantitative traits using RFLP linkage maps. Genetics. 121: 185-199.

176

LEAKEY, C.L.A. 1988. Genotypic and phenotypic markers in common bean. In: P.
Gepts (ed.). Genetic Resources of Phaseolus beans. Kluwer Academic
Publishers, Boston. p. 245-327.

LU, W. and KC. Chang. 1996. Correlations between chemical composition and canning
quality attributes of navy bean (Phaseon vulgaris L.). Cereal Chemistry. 73 (6):
785-787.

LU, W., K.C. Chang, K.F. Grafton and PB. Schwarz. 1996. Correlations between
physical properties and canning quality attributes of navy bean (Phaseolus
vulgaris L.). Cereal Chemistry. 73(6): 788-790.

McCLEAN, P., J. Ewing, M. Lince and K. Grafton. 1994. Development of a RAPD map
of Phaseolus vulgaris L. Ann. Rep. Bean Improvement Cooperative. 37: 79-80.

MEINERS, C.R., N.L. Derise, H.C. Lau, S.J. Ritchey and E.W. Murphy. 1976.
Proxirnate composition and yield of raw and cooked mature dry legumes. J.
Agric. Food Chem. 24:1122. as cited by S.K. Sathe, S.S. Despande and D.K.
Salunkhe. 1984. Dry beans of Phaseolus: A review. Part 1. Chemical
composition: proteins. CRC Crit. Rev. Food Sci. Nutr. 20: 1-46.

MELOTTO, M., R.A. Young and J .D. Kelly. 1998. Marker-assisted dissection of genes
conditioning resistance to anthracnose. Ann. Rep. Bean Improvement Cooperative.
41: 9-10.

MICHELMORE, R.W., I. Paran and RV. Kesseli. 1991. Identification of markers
linked to disease-resistance genes by bulked segregant analysis: a rapid method to

detect markers in specific genomic regions by using segregating populations.
Proc. Natl. Acad. Sci. USA. 88: 9828-9832.

MIKLAS, P. and J. Kelly. 1992. Identifying bean DNA polymorphisms using the
polymerase chain reaction. Ann. Rep. Bean Improvement Cooperative. 35: 21-22.

MIKLAS, P., E. Johnson, and J. Beaver. 1995. RAPD markers for QTLs expressing
BGMV resistance in dry bean. Ann. Rep. Bean Improvement Cooperative. 38:
11 1-1 12.

MIKLAS, P., E. Johnson, V. Stone, J .S. Beaver, C. Montoya and M. Zapata. 1996.
Selective mapping of QTL conditioning disease resistance in common bean. Crop
Sci. 36: 1344-1351.

MIKLAS, P.N., R. Delorrne, V. Stone, C.A. Urrea, J .8. Beaver and J .R. Steadman. 1998.

A RAPD map of disease resistance traits in common bean. Ann. Rep. Bean
Improvement Cooperative. 41:93-94.

177

MIKLAS, P.N., V. Stone, C.A. Urrea, E. Johnson and J .S. Beaver. 1998. Inheritance and
QTL analysis of field resistance to ashy stem blight in common bean. Crop Sci.
38: 916-921 .

NIENHUIS, J. and SP. Singh. 1985. Combining ability analyses and relationships
among yield, yield components and architectural traits in dry bean. Crop Sci. 26
(1): 21-27.

NIENHUIS, J. and SP. Singh. 1988. Genetics of seed yield and its components in
common bean (Phaseolus vulgaris L.) of Middle-American origin, 1. Genetic

variance, heritability and expected response from selection. Plant Breeding. 101:
155-163.

NODARI, R.O., S.M. Tsai, R.L. Gilbertson, and P. Gepts. 1993. Towards an integrated
linkage map of common bean. 2. Development of an RFLP-based linkage map.
Theor. Appl. Genet. 85: 513-520.

NORDSTROM, CL. and W.A. Sistrunk. 1977. Effect of type of bean, soak time,
canning media, and storage time on quality attributes and nutritional value of
canned dry beans. J. Food Sci. 42: 797—800.

NORDSTROM, CL. and W.A. Sistrunk. 1979. Effect of type of bean, moisture level,
blanch treatment and storage time on quality attributes and nutrients of canned dry
beans. J. Food Sci. 44: 392-395. as cited by W. Lu and KC. Chang. 1996.
Correlations between chemical composition and canning quality attributes of navy
bean (Phaseolus vulgaris L.). Cereal Chemistry. 73 (6): 785-787.

OGWAL, MO. and DR. Davis. 1994. Rapid rehydration methods for dried beans. J.
Food Science. 59(3):611-612, 654.

OLAYA, G., GS. Abawi and NP. Weeden. 1996. Inheritance of the resistance to
Macrophomina phaseolina and identification of RAPD markers linked to
resistance genes in beans. Phytopathology. 86(6): 674-679.

OSBORN, T.C., D.C. Alexander and J .F. Fobes. 1987. Identification of restriction
fragment length polymorphisms linked to genes controlling soluble solids content
in tomato. Theor. Appl. Genet. 73: 350-356.

PARAN, 1., R. Kesseli and R. Michelrnore. 1991. Identification of RFLP and RAPD
markers linked to downy mildew resistance genes in lettuce using near-isogenic
lines. Genome. 34: 1021-1027. as cited by J .E. Staub and RC. Serquen. 1995.
Genetic markers, map construction and their application in plant breeding.
HortScience. 31 (5): 729-740.

178

PARK, S.O., D.P. Coyne, G. Jung, E. Amaud-Santana and H. Ariyarathne. 1998.
Detection and mapping of RAPD markers associated with QTL affecting seed size

and shape in common bean. Ann. Rep. Bean Improvement Cooperative. 41: 147-
148.

PATERSON, AH. 1995. Molecular dissection of quantitative traits: progress and
prospects. Genome Research. 5: 321-333.

PATERSON, AH. 1998. Of blending, beans, and bristles: The foundations of QTL
mapping. In: Molecular dissection of complex traits. A.H. Paterson (ed.). New
York: CRC Press. pp. l-10.

QUAST, D.G. and SD. da Silva. 1977. Temperature dependence of hydration rate and
effect of hydration on the cooking rate of dry legumes. J. Food Sci. 42: 1299-
1303.

RANALLI, P., G. Ruaro and P. Del Re. 1991. Response to selection for seed yield in
bean (Phaseolus vulgaris). Euphytica. 57 (2): 117-123.

RIBAUT, J -M. and D. Hoisington. 1998. Marker-assisted selection: new tools and
strategies. Trends in Plant Science. 3 (6): 236-239.

RUENGSAKULRACH, S., N. Srisuma, M.A. Uebersax, G.L. Hosfield and LG. Occena.
Early generation screening of navy bean breeding lines by canning quality
assessment and pasting characteristics of bean flour. J. Food Quality. 17: 321-
333.

SARAFI, A. 1978. A yield-component selection experiment involving American and
Iranian cultivars of the common bean. Crop Sci. 18 (1): 5-7.

SATHE, S.K., S.S. Despande and D.K. Salunkhe. 1984. Dry beans of Phaseolus: A
review. Part 1. Chemical composition: proteins. CRC Crit. Rev. Food Sci. Nutr.
20: 1-46.

SAX, K. 1923. The association of size differences with seed coat pattern and
pigmentation in Phaseolus vulgaris. Genetics. 8: 552-560.

SCHABENBERGER, O. 1997. Statistics for Biologists 1. Course pack for STS/CSS
464, Section I: for Crop and Soil Sciences Majors. Fall Semester 1997. Michigan
State University, East Lansing, MI.

SCULLY, B.T., D.H. Wallace and DR. Viands. 1991. Heritability and correlation of

biomass, growth rates, harvest index, and phenology to the yield of common
beans. J. Amer. Soc. Hort.Sci. 116(1): 127.130. ‘

179

SHELLIE-DESSERT, KC. and RA. Bliss. 1991. Genetic improvement of food quality
factors. In: Common beans: Research for crop improvement. A.van
Schoonhoven and O. Voysest (eds.). CIAT: CAP International. pp. 649-677.

SIMMONDS, NW. 1979. Principles of crop improvement. New York: Longrnan, Inc.
408 pp.

SINGH, S.P., H. Teran, A. Molina and J .A. Gutierrez. 1991. Genetics of seed yield and
its components in common beans (Phaseolus vulgaris L.) of Andean origin. Plant
Breeding. 107(3): 254-257.

SINGH, SP. and CA. Urrea. 1995. Inter- and intraracial hybridization and selection for
seed yield in early generations of common bean, Phaseolus vulgaris L.
Euphytica. 81 (2) 131-137.

SKROCH, P.W., J .B. dos Santos and J. Nienhuis. 1992. Genetic relationships among
Phaseolus vulgaris genotypes based on RAPD marker data. Ann. Rep. Bean
Improvement Cooperative. 35: 23-24.

SKROCH, P., G. Jung, J. Nienhuis and D. Coyne. 1996. Integration of RAPD marker
linkage maps and comparative mapping of QTL for disease resistance in common
bean. Ann. Rep. Bean Improvement Cooperative. 39: 48-49.

STAUB, J .E., F.C. Serquen and M. Gupta. 1996. Genetic markers, map construction and
their application in plant breeding. HortScience. 31(5): 729-740.

STUBER, CW. and MD. Edwards. 1986. Genotypic selection for improvement of
quantitative traits in corn using molecular marker loci. 1986. Proc. 4lst Annual
Corn and Sorghum Research Con/C, Am. Seed Trade Assoc. 41: 40-83. as cited
by J .V. Dudley. Molecular markers in plant improvement: manipulation of genes
affecting quantitative traits. Crop Sci. 33: 660-668.

TANHUANPAA, P.K., J .P. Vilkki and H.J. Vilkki. 1996. Mapping of a QTL for oleic
acid concentration in spring turnip rape (Brassica rapa ssp. oletfera). Theor.
Appl. Genet. 92(8): 952-956.

TANKSLEY, SD. and J .C. Nelson. 1996. Advanced backcross QTL: a method for the
simultaneous discovery and transfer of valuable QTLs from unadapted germplasm
into elite breeding lines. Iheor. App. Genet 92: 899-909.

TIMMERMAN, G.M., T.J. Frew, N.F. Weeden, A.L. Miller and D.S. Goulden. 1994.
Linkage analysis of er-I , a recessive Pisum sativum gene for resistance to
powdery mildew fungus (Erysiphe pisi D.C.). Theor. Appl. Genet. 88: 1050-1055.
as cited by J .E. Staub and RC. Serquen. 1995. Genetic markers, map
construction and their application in plant breeding. HortScience. 31 (5): 729-740.

180

UEBERSAX, M.A. 1972. Effects of storage and processing parameters on quality
attributes of processed navy beans. M.S. Thesis, Michigan State University.

UEBERSAX, M.A. and C.L. Bedford. 1980. Navy bean processing: effect of storage
and soaking methods on quality of canned beans. Mich. State Univ. Agr. Expt.
Sta., E. Lansing Res. Rpt. 410.

UEBERSAX, M.A. 1985. Quality aspects of moisture, soaking and blanching in dry
bean processing. In: Proc. of Tech. Conf. on Dry Bean Research. San Francisco,
February 13, Food Processors Institute, Washington DC p. 7. as cited by KL.
Wiese and ER. Jackson. 1993. Changes in thermal process time (B1,) for baked
beans based on water hardness and fill temperature. J. Food Protection. 56(7):
608-611.

UEBERSAX, M.A., S. Ruengsakulrach and LG. Occena. 1991. Strategies for
processing dry beans. Food Technol. 45:104. as cited by MD. ngal and DR.
Davis. 1994. Rapid rehydration methods for dried beans. J. Food Science.
59(3): 611-612, 654.

USDA (United States Department of Agriculture). 1982. The United States standards for
beans. Federal Grain Inspection Service. US. Department of Agriculture. 16 p.
as cited by O. Voysest and M. Dessert. 1991. Bean cultivars: Classes and
commercial seed types. In: Common beans: Research for crop improvement. A.

van Schoonhoven and O. Voysest (eds.). CAB International: CIAT. pp.119-162.

USDA-NASS (United States Department of Agriculture - National Agricultural Statistics
Service). 2000. Agricultural statistics 2000. US. Government Printing Office,
Washington. at http:l/www.usda.gov/nass/pubs/agstats.htm.

USDA-NASS (United States Department of Agriculture - National Agricultural Statistics
Service) and Michigan Department of Agriculture. 2000. Michigan Agricultural
Statistics 1998-1999. at http://www.mda.state.rni.us/mass/statsOO/Crops00.htm.

USDA-NASS (United States Department of Agriculture - National Agricultural Statistics
Service) and Minnesota Department of Agriculture. 2000. Minnesota
Agricultural Statistics 2000. at http://www.nass.usda.gov/mn.

USDA-NASS (United States Department of Agriculture - National Agricultural Statistics
Service) and North Dakota Department of Agriculture. 2000. North Dakota
Agricultural Statistics 2000. at http://www.nass.usda.gov/nd.

VAN BUREN, J ., M. Boume, D. Downing, D. Queale, E. Chase and S. Comstock. 1986.

Processing factors influencing splitting and other quality characteristics of canned
kidney beans. J. Food Science. 51(5): 1228-1230.

181

WALTERS, KJ. 1995. Identification of RAPD markers associated with canning quality
in navy beans. M.S. thesis. Michigan State University.

WALTERS, K.J., G.L. Hosfield, M.A. Uebersax and J .D. Kelly. 1997. Navy bean
canning quality: correlations, heritability estimates, and randomly amplified
polymorphic DNA markers associated with component traits. J. Amer. Soc. Hort.
Sci. 122(3): 338 - 343.

WANG, OCR. and S.K.C. Chang. 1988. Effect of selected canning methods on trypsin
inhibitor activity, sterilization value, and firmness of canned beans. J. Agric.
Food Chem. 36: 1015-1018.

WANG, C.R., K.C. Chang and K. Grafton. 1988. Canning quality evaluation of pinto
and navy beans. J. Food Science. 53 (3): 772-776.

WANG, G.L. and AH. Paterson. 1994. Assessment of DNA pooling strategies for
mapping of QTLS. Theor. Appl. Genet. 88: 355-361.

WASSIMI, N.N., G.L. Hosfield and M.A. Uebersax. 1990. Inheritance of physico-
chemical seed characters related to culinary quality in dry bean. J. Amer. Soc.
Hort. Sci. 115:492-499.

WEEDEN, N.F., M. Timmerrnan, M. Hermmat, B.E. Kneen and M.S. Lodhi. 1992.
Inheritance and repeatability of RAPD markers. In: J. Nienhuis (ed.). Proc.
Symp. Applications of RAPD Technology in Plant Breeding, 12-17. Nov. 1992.
Minneapolis, MN. as cited by J .E. Staub and F .C. Serquen. 1995. Genetic
markers, map construction and their application in plant breeding. HortScience.
31 (5): 729-740.

WELSH, W., W. Bushuk, W. Roca and SP. Singh. 1995. Characterization of agronomic
traits and markers of recombinant inbred lines from intra- and interracial
populations of Phaseolus vulgaris L. Theor. Appl. Genet. 91 (1) 169-177.

WIESE, KL. and ER. Jackson. 1993. Changes in thermal process time (B1,) for baked
beans based on water hardness and fill temperature. J. Food Protection. 56(7):
608-611.

WILLIAMS, J.G.K., A.R. Kubelik, K.J. Livak, J .A. Rafalski and S.V. Tingey. 1990.
DNA polymorphisms amplified by arbitrary primers are useful as genetic markers.
Nucleic Acids Research. 18 (22): 6531-6535.

WYCH, R.D., R.L. McGraw, and DD. Stuthman. Genotype x year interaction for length
and rate of grain filling in oats. Crop Sci. 22: 1025-1028.

182

XIAO, J ., J. Li, S. Grandillo, S.N. Ahn, L. Yuan, S.D. Tankley and SR. McCouch. 1998.
Identification of trait-improving quantitative trait loci alleles fiom a wild rice
relative, Oryza rufipogon. Genetics. 150: 899-909.

YAN, W. and DH. Wallace. 1995. Breeding for negatively associated traits. Plant
Breeding Reviews. 13: 141-177.

YOUNG, R.A. and J .D. Kelly. 1996. Gene pyramiding using marker assisted selection
for stable resistance to bean anthracnose. Ann. Rep. Bean Improvement
Cooperative. 39: 57-58.

ZIMMERMAN, M.J.O., A.A. Rosielle and J .G. Waines. 1984a. Heritabilities of gain
yield of common bean in sole crop and intercrop with maize. Crop Sci. 24(4):
641-644.

ZIMMERMAN, M.J.O., A.A. Rosielle, J .G. Waines and K.W. Foster. 1984b. A
heritability and correlation study of gain yield, yield components, and harvest
index of common bean in sole crop and intercrop. Field Crops Res. 9: 109-118.

ZIMMERMAN, M.J.O., A.A. Rosielle, K.W. Foster and J .G. Waines. 1985. Gene

action for gain yield and harvest index of common bean gown as sole crop and
in intercrop with maize. Field Crops Res. 12(4): 319-329.

183