John Milton Poehlman, David Sleper Breeding Fiel (B Ok - Org) - 169-249
John Milton Poehlman, David Sleper Breeding Fiel (B Ok - Org) - 169-249
159
9. Breeding SelfPollinated Crops
Before discussing the methods by which new cultivars of selfpollinated crops originate, let us consider the question, ''What is a cultivar?"
What is a Cultivar?
The cultivar (agricultural variety) is a group of genetically similar plants, which by structural features and performance may be identified from other groups of
genetically similar plants within a species. The cultivar as an agronomic unit is familiar to plant breeders who develop new cultivars and to seedsmen and farmers who
multiply the seed and grow the cultivars. But to understand the concept of the cultivar requires a knowledge of the system of plant classification. The plant kingdom is
divided into taxonomic groups of similar and closely related plants; in this scheme, families of plants are divided into genera, which are subdivided into species.
Within the species, the agronomist and horticulturalist recognize numerous agricultural varieties, more commonly referred to as varieties or cultivars.
This relationship can be clarified by examining the taxonomic classification of a common crop plant, the soybean, a species in the family, Leguminosae:
Family: Leguminosae (subfamily Papilionoideae)
Genus: Glycine
Species: max
The scientific name of the cultivated soybean is Glycine max; the first word designates the genus, the second word the species. The species, G. max, contains many
forms that are genetically different and distinguished from each other by heritable traits such as maturity, seed color, pubescence color, plant type, disease resistance,
oil content of seed, and a host of other characteristics.
A population of soybean may be composed of a single genotype or a mixture of genotypes and may be variously referred to as an experimental strain, a strain, or a
line. Thousands of
Page 160
experimental strains are generated in the plant breeder's nursery each year. Once a superior strain is identified, it may be named, the seed increased, and distributed as
an 'agricultural variety' or 'cultivar'. Earlier, the term 'variety' was commonly used by farmers and seed producers; later the term 'cultivar' was coined to serve as the
international equivalent of a cultivated variety. Variety and cultivar may be used interchangeably, but cultivar is now preferred in scientific literature and is used in this
text. The distinction of being named and distributed commercially serves to set apart the cultivar from the experimental strain or breeding line. In the United
States, the name, description, and developer of new field crop cultivars are registered by the Crop Science Society of America and this information is published in
Crop Science.
Two essential characteristics of a cultivar are (1) identity and (2) reproducibility. Identity is necessary so that the cultivar may be recognized and distinguished from
other cultivars within the species. Typically, the distinguishing features may be morphological structures, color markings, physiological response, disease reaction, or
performance. Reproducibility is needed so that the characteristics by which the cultivar is identified will be reproduced in the progeny. In selffertilized crops, a cultivar
increased from a single, homozygous genotype will be uniform in appearance, whereas a cultivar increased from a mixture of genotypes will exhibit a range of genetic
variability according to that present in the mixture.
Genetic Significance of Pollination Method
Selfpollinated crops differ in genetic makeup from plants in crop species that are normally crosspollinated. In a crop that is selfpollinated, it is the rule that plants
will be homozygous. This assumption may be made because:
• loci with identical genes (AA or aa) will remain homozygous following selfpollination,
• loci with contrasting genes (Aa) will segregate, producing homozygous and heterozygous progeny in equal proportions.
Heterozygosity is reduced by 50% with each successive selffertilization(Fig. 9.1). After several successive generations of selfpollination, the proportion of
heterozygous loci remaining in a population is very small. Although complete homozygosity is theoretically unattainable, plants selected from a mixed population after
five to eight generations of selfing will normally have reached a practical state of homozygosity such that their progeny will be uniform in appearance and performance.
Breeding procedures in selfpollinated species are based on the genetic structure of selfpollinated populations. A mixed population of a selfpollinated crop is
composed of plants with different homozygous genotypes. If single plants differing in genotype are harvested and the seed increased, each will produce a pure
population, although the populations will differ from each other. Heterozygous plants may arise in a population of a selfpollinated crop through (1) crosspollination
among plants with different genotypes, or (2) by mutation. The progenies of the heterozygous plants will quickly segregate in succeeding generations giving rise to
homozygous subpopulations.
Page 161
Fig.9.1.
Proportions of homozygous and heterozygous genotypes in a population after
successive generations of selfpollination, assuming equal fitness for survival
among genotypes. S0, original selfed plant; S1, first selfed generation; S2,
second selfed generation; and so on.
Breeding Methods in SelfPollinated Crops
A new cultivar of a selfpollinated crop normally originates from an increase of:
• a mixture of plants, or a single plant, selected from introduced germplasm,
• a mixture of plants, or a single plant, selected from a local population, or
• a single plant selected from a hybrid population.
Increases from introduced germplasm or local populations involves the identification and increase of established genotypes, whereas selection from a hybrid population
involves the creation of new genotypes through crosspollination and gene recombination. Most new cultivars of selfpollinated crops are developed by the latter
procedure. In the breeding of selfpollinated crops, thousands of strains are normally grown in adjacent plots in the breeding nursery without pollination control. Some
natural crosspollination generally occurs, but the amount is usually so small that it is ignored except when extreme purity is essential, as in genetic studies, or when a
strain is being increased for final distribution as a new cultivar. If the amount of natural crosspollination is sufficient to visibly affect uniformity, reselection is practiced
to repurify the strain.
Assembly of Germplasm
The initial step in a breeding program is to assemble a wide assortment of germplasm (genetic strains of diverse origin) of the desired species, always searching for
accessions with genes that will contribute to improved performance. Commercial cultivars are a desirable source of useful germplasm, except where their use is
restricted by legal protection. Advanced breeding lines with proven adaptation and productivity are another useful source. Usually, the latter may be assembled from
state, national, or international breeding programs or from
Page 162
genebanks. In earlier years, cultivated native varieties, called landraces, could occasionally be introduced and utilized as new cultivars. Today, the landraces have
been displaced in cultivation by modem cultivars except in the most primitive areas of the world. Landraces still serve as a source of useful genes in crop breeding
programs, but they are seldom available except from genebanks around the world where they have been collected and stored to provide useful genes for future use by
plant breeders. Germplasm accessions should be grown initially in the local environment to identify sources of genes for maturity, yield potential, disease resistance,
and other desired traits, and to observe inherent weaknesses.
Selection
Selection, as a breeding procedure, involves identification and propagation of individual genotypes or groups of genotypes from mixed populations, or from
segregating populations following hybridization. Unless genetic variation can be identified and distinguished from environmentally caused variability within the mixed
population, selection may not be effective in isolating the desired genotypes. Selection procedures practiced in mixed populations of selfpollinated crops are mass
selection and pureline selection. The populations created are referred to as mass selections or pure lines, respectively.
MASS SELECTION. In the massselection procedure, plants are chosen and harvested on the basis of phenotype and the seeds composited without progeny
testing. Cultivars developed by mass selection are normally uniform for qualitative characters with simple genetic inheritance, such as presence of awns, color
markings, or maturity, where phenotypic differences can be visibly recognized and utilized as selection criteria. Genetic variations in quantitative traits such as yield,
size, or quality, where phenotypic differences are too small to be recognized, or cannot be accurately distinguished from environmentally caused variations, may still be
present.
The objectives in mass selection are to:
• purify a mixed cultivar or plant population by selecting and propagating visibly similar plants, or
• develop a new cultivar by improving the average performance of the population.
If a mixed cultivar or plant population is purified by mass selection, testing may be terminated and seed increase started any time after it has been verified that the new
strain does not differ in adaptation and performance from the original population, and that it is superior to the original population in uniformity. Some genetic variation
within a mass selection may be useful by providing buffering against variations in the environment. With mass selection, it is not possible to distinguish between plants
that are homozygous or heterozygous for a qualitative character controlled by a dominant gene. The heterozygous plants will segregate in the following generation and
phenotypic selection may need to be repeated. Neither is it possible to distinguish whether a plant owes its superior appearance to hereditary characters or to
favorable environmental influences.
PURELINE SELECTION. A pure line is a progeny descendent solely by selfpollination from a single homozygous plant. Pureline selection refers to the
procedure of isolating pure lines from a mixed population. A cultivar developed by pureline selection is more uniform than a cultivar developed by mass selection,
because all of the plants in the cultivar will have
Page 163
the same genotype. This is assuming that the plant originally selected is homozygous at all loci, an assumption plant breeders often make, but a condition seldom if ever
achieved.
In the past, pureline cultivars were often developed by identifying superior plants in a landrace, or a mixed cultivar. Currently, most new cultivars originate from
selections made in segregating progenies following hybridization. In either case, the progeny test, as described in Chapter 3, is an essential feature of pureline
selection and serves to evaluate the breeding behavior of the selected plant. Pureline selection does not create a new genotype, and improvement is limited to the
isolation of the best genotype present in the mixed population. Once the superiority of a selected genotype has been proven, the population may be increased, named,
and distributed as a new cultivar.
How long does the new cultivar remain pure? That depends upon amount of:
• seed mixtures from other sources,
• natural crossing with other cultivars or breeding lines, and
• mutations.
Harvesting and seedcleaning equipment are common sources of seed mixtures, either in the breeding nursery, or with commercial lots of seed after cultivar release.
Natural crossing may occur between plants of the cultivar and plants occurring as mixtures within the cultivar. The amount of natural crossing will vary with the crop
species but rarely exceeds 1 to 2 % in selfpollinated crops. Offtype plants resulting from natural crossing or mutation need to be rogued out to maintain cultivar
purity.
THE PURELINE THEORY. The theory of the pureline was established by the Danish botanist, Johannsen, in 1903. Johannsen conducted selection experiments
for seed weight in a mixed seed lot of the 'Princess' bean (see Fig. 4.3). Because beans are selffertilized, the seeds in the original lot were homozygous for genes
affecting seed weight. Selection within the original mixed lot of beans was effective in isolating lines that were genetically different. Once the pure line was isolated,
further selection within the pure line was ineffective. In Johannsen's original mixed lot of beans, the variations in seed weight were both hereditary and environmental;
within the pure lines, the variations were due only to differences in the effects of the environment.
PURE LINES VS. GENETIC DIVERSITY. In the selfpollinated crops, major attention has been given to the development of uniform, pureline cultivars.
Uniformity in appearance has a cosmetic value sometimes referred to as "eyeappeal," uniformity in maturity facilitates harvesting, and uniformity in quality enhances the
market value of the product. In recent years the opinion has been growing that cultivars with more genetic variability would produce stable yields over a greater range
of environments and seasons, be more widely adapted, and offer broader protection against specialized disease producing organisms than cultivars with limited genetic
variability. An objection to the genetic diversity is based on the grounds that cultivar identification in variety protection and seed certification programs becomes more
difficult and less accurate.
Hybridization
Hybridization is a breeding method that utilizes crosspollination between genetically different parents to obtain gene recombination. Following the crosspollination,
segregating
Page 164
generations are grown, and pure lines selected after homozygosity is reached. The objective is to identify and select lines which combine desirable genes from both
parents. The selected lines are evaluated by progeny tests to verify presence of a desirable combination of genes. Lines proven to be superior may be increased as a
new cultivar. Hybridization breeding should not be confused with the breeding of hybrid cultivars in which a cross is made and the F1 generation is grown.
In addition to combining genes for visible traits of the parents, plants may be selected from the segregating progenies that fall outside the range of the parents. Those
falling outside the range of the parents are known as transgressive segregates. Transgressive segregates with a combination of genes superior to either parent in a
quantitative feature, such as yield, seed weight, winter hardiness, or straw stiffness, in which inheritance is determined by multiple genes, may be selected and utilized in
the breeding program.
In a cross between pureline parents, all F1 plants will have identical genotypes, and be heterozygous at loci where the parents have contrasting alleles. Genetic
segregation begins and maximum segregation occurs in the F2 generation, with heterozygosity reduced by 50 % with each succeeding selfed generation (Fig. 9.1). The
number of F1 plants to grow will depend upon the desired size of the F2 progeny. Usually a large F2 population, from 1000 to 10,000 plants, will be needed to provide
the desired range of genetic segregation; the number will be affected by how many traits the parents have in common and the number of different genes from each
parent that the breeder desires to combine in the progeny. After segregation has virtually ceased, (5th or 6th generation), plants with a superior combination of the
desired parent characteristics need to be identified and increased as a pure population. Performance of the new lines are evaluated in field trials in comparison with the
parent lines.
Selection Procedures Following Hybridization
Selection procedures that may be used to identify desirable genotypes from segregating progenies, following hybridization in selfpollinated crops, include (1)
pedigreeselection, (2) bulkpopulation, (3) singleseeddescent, and (4) doubledhaploid.
PedigreeSelection
In the pedigreeselection procedure, selection for plants with the desired combination of characters is started in the F2 generation, and continued in succeeding
generations until genetic purity is reached (Fig. 9.2). An example of the pedigreeselection method follows:
Crossing generation. Cross cultivar A × cultivar B.
F1 generation. Grow 50 to 100 F1 plants. Before harvest, eliminate plants that may have arisen from selfpollination.
F2 generation. Grow 2000 to 3000 F2 plants. Space plants sufficiently that individual plants may be examined. Select and harvest superior plants in which desired
characteristics of the parent cultivars are combined, harvesting the seed separately from each plant.
F3to F5 generations. Grow progeny rows with seed harvested from superior plants harvested in the previous generation. Space plants in the row so that individual
plants may be studied. Identify superior rows, then select and harvest 3 to 5 of the best plants within these rows. Continue selection between and within rows through
the F5 generation. Normally, 25 to
Page 165
Fig. 9.2.
Pedigree method of selection. From selected F2 plants, progenies of 25 to 30 plants
are grown in plant rows in F3. Superior plants from the best rows are selected and
planted in families of plant rows in F4 to F6, with selection being made of best plants,
in best rows, of best families. By F6 families should be relatively uniform. Preliminary
yield trials are planted in F7, and yield trials are continued through F10. Various
modifications of this procedure may be made. For example, after plants are selected in F3
and F4, remaining plants in row may be bulked and preliminary yield tests started.
50 families may be retained at the end of the F5 generations. Identity of plant and row is maintained and superior traits of the plants are recorded.
F6 generation. Grow families of plant rows. Uniform related families may be harvested together and the seed bulked. The separate seed lots are designated
experimental lines.
F7 generation. Grow the experimental lines in a preliminary yield trial in comparison with adapted cultivars.
F8 to F10 generations. Yield trials of superior experimental lines are continued at two or more locations in comparison with adapted commercial cultivars. Only the
highest yielding lines are retained for testing in the next yield trial. During the testing period, observations are made on height, tendency to lodge, maturity, disease and
insect resistance, quality, and other characteristics as appropriate in the crop being studied. Growing the lines in regional yield
Page 166
trials in environmentally diverse locations will assist in identifying lines with adaptation to a wide range of environments. If, after 3 to 5 years of yield testing, lines
superior to the check cultivars have been identified, one line may be chosen for increase and distributed as a new cultivar.
F11 and F12 generations. Increase seed and distribute the new cultivar.
Modifications of the pedigreeselection procedure may be employed, such as introducing yield trials as early as the F3 or F4 generation. Only the highyielding lines are
then grown in advanced generations. Or, selection may be terminated earlier than indicated if the lines appear to be uniform.
The pedigreeselection method is labor intensive and requires detailed recordkeeping during the early segregating generations. It has the advantage that only progeny
lines in which plants with genes for the desirable characters have been identified are carried forward to the next generation. This method also permits the collection of
genetic information which is not possible with other procedures. The pedigreeselection method of breeding is best suited to crops where individual plants may be
examined and harvested separately, as in cereals, garden bean, peanut, soybean, tobacco, or tomato. Although widely used in particular species, the pedigree system
is not always pursued as fully as described due to the labor requirements. Resources may be spread by growing more progeny rows, with a relaxation in notetaking
and recordkeeping. As illustrated, the pedigreeselection procedure requires 12 years to develop a cultivar if one generation only is grown each year. The number of
years may be reduced by growing more than one generation per year, either in the greenhouse or by growing winter or offseason nurseries in an area with a favorable
climate. The range of natural selection may be increased if the segregating generations are grown in more than one environment.
BulkPopulation
In the bulkpopulation procedure, seeds harvested in the F2 and succeeding generations are bulked and grown, with selection delayed until an advanced generation,
commonly the F5 or the F6, at which time the segregation will have virtually ceased (Fig. 9.3). An example of the bulkpopulation procedure follows:
Crossing generation. Cross cultivar A × cultivar B.
F1 generation. Grow 50 to 100 F1 plants. Before harvest, eliminate plants that may have arisen from selfpollination. Harvest en masse and bulk seed.
F2 generation. Grow 2000 to 3000 F2 plants. Harvest en masse and bulk seed from all plants.
F3 to F4 generations. Grow 1/50 to 1/100hectare plots with bulked seed harvested from the preceding generation.
F5 generation. Space plant 3000 to 5000 seeds. Select and harvest 300 to 500 superior plants keeping seed separate from each plant.
F6 generation. Grow progeny rows of selected plants; harvest 30 to 50 progenies in which plants exhibit the desired characteristics of the parents.
F7 generation. Grow superior progenies harvested in the F6 in a preliminary yield trial.
F8 to F10 generations. Yield trials are continued in multiple locations as in the pedigreeselection procedure.
F11 and F12 generations. Increase seed of a superior line and distribute as a new cultivar.
Page 167
Fig. 9.3.
Bulkpopulation method of selection. The progeny of the cross is grown in
a bulk planting through the F4 generation. In F5 the progeny is space planted.
Plant or head selections are made and grown in plant or head rows in F6. Superior
rows are selected and grown in a preliminary yield trial in F7. Superior strains are
grown in yield trials in F8 to F10. Various modifications of this procedure may be
made. For example, selection may start as early as F3 or F4, with lines having a
superior yield being purified in later generations, or bulk plots may be replicated
and harvested for yield and entire crosses discarded on the basis of the yield
of the bulk plots.
The bulkpopulation method of breeding is simple, convenient, requires less labor, and is less expensive to conduct during the early segregating generations than the
pedigreeselection procedure. It is necessary to grow large populations of spaced plants in the selection generation to have a reasonable chance of finding desirable
segregates. Subjecting the bulkpopulations to disease epidemics, winter injury, drought, or other adversities during the segregating
Page 168
generations will foster natural selection for those features. Lines selected from the bulkpopulation that are still segregating will need to be reselected to establish true
breeding strains. The bulkpopulation method is suited to crops normally planted in thick spacings, like small grains, in which it is difficult to separate and identify
individual plants. In contrast to the pedigreeselection method, no information is obtained during the early generations on inheritance of specific traits or performance of
specific lines. During the segregating generations some desirable genotypes may be lost from the population, for example, tall and late plants may suppress short and
early plants. The bulkpopulation selection procedure may be modified by selecting in the F3 or the F4 and starting yield trials even though the lines are still segregating.
Superior yielding lines may be reselected while yield testing continues.
SingleSeedDescent
In singleseeddescent the progenies of the F2 plants are advanced rapidly through succeeding generations from single seeds (Fig. 9.4). An example of the singleseed
descent procedure follows:
Crossing generation. Cross cultivar A × cultivar B.
F1 generation. Grow 50 to 100 F1 plants.
F2 generation. Grow 2000 to 3000 F2 plants. Harvest a single seed from each plant. Identity of the F2 plant is not maintained.
F3 and F4 generations. Grow seeds harvested in previous generation. Harvest a single seed from each plant.
F5 generation. Space plants in field from seeds harvested in previous generation. Select plants superior for desired characteristics and harvest seeds from the selected
plants.
F6 generation. Grow progeny rows from plants harvested in the previous generation. Harvest rows superior for desired characteristics. Each row will have originated
from a different F2 plant.
F7 generation. Grow preliminary yield trial from rows harvested in the previous generation.
F8 to F10 generations. Continue yield trials in multiple locations as in pedigreeselection and bulkpopulation procedures.
F11 and F12 generations. Increase superior line and distribute as a new cultivar.
An alternative procedure would be to space plant the F4 generation and plant the F5 in rows, thereby getting lines into yield trials one generation earlier.
The singleseeddescent procedure was proposed as a means of maintaining descendants from the maximum number of F2 plants, thereby reducing the loss of
genotypes during the segregating generations. As currently practiced, singleseed descent is utilized to reduce the time required to grow the segregating generations.
Because only one seed is harvested from each plant, optimum plant development in the F2 to F4 generations is unnecessary. By thickly planting seeds in a greenhouse
bench, growing plants with low soil fertility, and using temperature and lighting regimes that force early flowering, two to three generations are commonly harvested in a
1year period, and the preliminary yield trial can be reached 1 to 2 years earlier. Species that can be forced to mature rapidly, such as soybean or summergrown
cereals (wheat, oat, barley), are suited for the singleseeddescent procedure. For winter cereals in which a vernalization period is required to obtain flowering, the
efficiency of the
Page 169
Fig. 9.4.
Singleseeddescent method of selection. Seeds harvested from F1 plants are
space planted in F2. A single seed harvested from each F2 plant is used
to plant the F3. Succeeding generations through the F5 are likewise planted
from single seeds harvested from each plant grown in the preceding generation.
In the F5 generation plants are harvested and a progeny row grown in the F6.
A preliminary yield trial is grown in the F7 and yield trials continued through the F10.
Some breeders combine singleseed descent with the pedigree selection procedure,
by growing only the F3 and F4 by singleseed descent to accelerate the time required
to reach a yield trial.
Page 170
singleseeddescent procedure is reduced.
A modification of the procedure has developed in soybean in which several pods are harvested from each plant instead of a single seed. A two to threeseed sample
from each plant is bulked to grow the next generation. This modified singleseeddescent procedure has become the principal method of advancing early generations in
soybean breeding programs. With tropical winter nurseries, two or three generations can be advanced in a year with the modified procedure.
With singleseeddescent, weak plants are not eliminated as in a fieldgrown nursery, and there is no provision for selection of superior segregates within families
descendent from F2 plants. Modifications to the procedure may be introduced, such as screening for disease resistance or other appropriate characteristics in any
generation. No recordkeeping is required during the early segregating generations. Final evaluation of progenies and yield trials are conducted in the field.
DoubledHaploid
In the doubledhaploid procedure, haploid plants are generated from anthers of F1 plants, or by other means, and the chromosomes of the haploid plants are doubled
with colchicine to produce diploid plants (Fig. 9.5). An example of the doubledhaploid procedure using anther culture follows:
Crossing generation. Cross cultivar A × cultivar B.
F1 generation. Culture anthers to produce 2000 to 3000 haploid plants.
F2 generation. Double chromosomes of haploid plants and harvest seeds from the doubledhaploid plants produced.
F3 generation. Grow progeny rows from doubledhaploid plants and harvest seed from superior rows.
F4 generation. Grow progeny rows in the field and select superior lines.
F5 generation. Grow preliminary yield trial.
F6 to F8 generations. Continue yield trials.
F9 and F10 generations. Increase and distribute superior line as a new cultivar.
Doubledhaploid plants are normally homozygous at all loci and it is unnecessary to grow segregating generations. Lines generated by the doubledhaploid procedure
may reach preliminary yield trials two to three generations earlier than with the pedigreeselection or bulkpopulation procedures. Like the singleseeddescent
procedure, early generations are not exposed to environmental stresses in the field, and attrition of lines is greater in initial field evaluation trials than with pedigree
selection or bulkpopulation procedures, in which the early generations are field grown. For successful use of the doubledhaploid procedure in plant breeding,
efficient and reliable techniques for generating haploid and doubledhaploid plants are essential. The doubledhaploids should be vigorous, stable, free from tissue
cultureinduced variations, and represent a random selection of the F1 pollen gametes. Current procedures for production of haploids and doubledhaploids have only
been partially successful in attaining these characteristics.
Choice of Procedure
The superiority of a new cultivar is determined by the combination of genes that it contains, not by the procedure by which the cultivar was produced. Choice of
procedure should
Page 171
Fig. 9.5.
Doubledhaploid procedure. Crosses are made and F1 progeny grown as in
previous procedures. Anthers from F1 plants are cultured and chromosome number
of haploid plants generated is doubled with colchicine to produce doubled haploids.
Progenies of doubledhaploid plants are evaluated in the field in F3 and F4 and superior
lines tested in yield trials in F5 to F5 generations.
be determined by the efficiency with which a superior combination can be assured and will vary with the crop species, the breeding objectives, and the resources
available to the breeder. The current trend is to adopt breeding procedures that will reduce the number of years to develop and release a cultivar, and that will enable
the breeder to grow and examine the largest number of lines with the resources available. Reduction in years may be accomplished by
Page 172
growing two or more generations per year through winter nurseries, by singleseeddescent, or by reducing the number of segregating generations as with the doubled
haploid procedure.
The pedigreeselection procedure is certainly the most precise system when the objective is to combine particular parent traits that are simply inherited and easily
observed in progeny plants, but it is less precise if the characters to be combined are quantitatively inherited, particularly if the heritability is low. It is laborintensive in
the early generations, due to the extensive seed packaging, planting, notetaking, and recordkeeping required. The bulkpopulation system has gained favor due to
the economy of labor and ease of growing large populations in the early generations. Singleseeddescent is suitable for crops that can be grown in a greenhouse
environment or in winter nurseries in a semitropical climate. It is economical to pursue, and reduces the time required to grow the early segregating generations. The
doubledhaploid procedure is labor intensive in the production of haploids and does not have the proven reliability of the other procedures.
Success in the hybridization method of breeding selfpollinated crops is dependent upon:
• choosing the correct parents, and
• identifying the superior plants from the segregating populations.
The choice of parents will be facilitated by clear and specific breeding objectives and superiority of the parents in characteristics contributing to those objectives. The
contributions from the parents should complement each other, so that selected progeny plants will not be lacking in some important agronomic characteristic.
Identification of the superior genotypes in the segregating progenies requires exhaustive testing and exposure to many adversities (e.g., disease, drought, or cold),
extensive observation in various stages of growth, and accurate recording of the observations. Testing in different seasons at several locations with diverse climatic
conditions will aid in identifying genotypes adapted over wide geographic areas. Only those lines that are distinctly superior and fulfill the objectives of the cross should
be propagated, with rigorous rejection of mediocre selections or crosses. The latter requires judgment decisions that can best be made by a skilled and
experienced breeder.
If only one quantitative character is being emphasized in a cross, it should be possible to select transgressive segregates superior to either parent. If two or more
quantitative characters are being improved, some compromise may be necessary, because one would seldom find superior transgressive segregation occurring
simultaneously for two or more characters. At this point the breeder must choose which line best exemplifies the objectives of the cross and will be increased for
further evaluation.
Backcross Breeding
The backcross is a form of recurrent hybridization by which a desirable allele for a character is substituted for the alternative allele in an otherwise desirable cultivar.
The plan of the backcross is to cross an adapted and productive cultivar, yet one that lacks a desirable allele (or alleles) controlling a superior character, to a breeding
line or cultivar in which the desirable allele is present. Beginning in the F1 and continuing for several generations, hybrid plants containing the dominant allele are
selected and successively crossed back to the adapted parent cultivar. The adapted parent, to which the allele is being added, enters into each backcross and is called
the recurrent parent. The parent with the superior character enters into the initial cross
Page 173
but does not enter into the backcrosses, and is called the donor or nonrecurrent parent.
The purpose of the backcross is to recover the genotype of the recurrent parent, except for the substitution of the allele (or alleles) for superior expression of the
character being contributed from the donor or nonrecurrent parent. The backcross is a form of inbreeding, and the features of the recurrent parent are automatically
recovered after successive backcrosses. The only selection practiced is for the one superior trait contributed by the nonrecurrent parent. The number of backcrosses
may vary from two to five, or more, depending upon how completely the breeder wishes to recover the genes from the recurrent parent. The backcross procedure is
most easily carried out if the character being added is simply inherited, dominant, and easily recognized in the hybrid plants.
The backcross procedure is illustrated in Fig. 9.6 in which the dominant alleles for a gene controlling disease resistance (RR) are to be substituted for the recessive
alleles in an adapted cultivar. In this cross cultivar A is the recurrent parent, and contains the genes for adaptation and yield that the breeder wishes to recover in the
new cultivar. Cultivar B is the donor parent with a dominant allele for disease resistance that the breeder wishes to add to cultivar A. With each successive backcross,
the progeny becomes more like the recurrent parent as additional genes for adaptation are recovered. With completion of the fourth backcross, theoretically, 93.75 %
of the genes of the adapted parent will have been recovered in the backcross progeny. After each backcross, disease resistant progeny plants (Rr) are identified, by
artificially inoculating all progeny plants and noting their disease reaction. As many backcrosses may be made as are necessary to obtain plants that are
indistinguishable from the recurrent parent except for the substituted allele for disease resistance. The diseaseresistant plants in the final backcross progeny will be
heterozygous for resistance (Rr), and must be selfed for one generation to obtain true breeding resistant plants (RR).
If the alleles for disease resistance being transferred should be recessive (rr), the progeny of the first backcross would segregate into genotypes (RR) and (Rr).
Because the heterozygous plants that contain the resistance allele (r) cannot be identified, it is necessary to self the progeny one generation to find resistant (rr) plants
before making the next backcross to the recurrent parent. Another procedure would be to backcross both the homozygous (RR) and heterozygous (Rr) plants to the
recurrent parent and simultaneously self each backcross derived plant and test the selfed progenies for resistance. The backcross progenies from the plants that prove
to be heterozygous are then kept, and the backcross progenies from the homozygous plants are discarded. If genes for undesirable characters are closely linked with
the gene for resistance, they may be added along with the resistance gene. The new cultivar would then differ from the recurrent parent by the genes that were added.
If characteristics being added by the backcross procedure are determined by multiple genes, it will be necessary for the backcross progenies to be grown through the
F2 or later generations to obtain plants that exhibit the desired characteristics before proceeding with the next backcross.
One feature of the backcross procedure is that a backcross derived cultivar will be adapted in the same general environment as the recurrent parent, reducing the
testing normally necessary to confirm adaptation of the backcross derived cultivar. An additional feature is that it is repeatable. A breeder can recover the same line if
the same recurrent and donor parents are used. If two or more characters are to be added to the same recurrent cultivar, separate backcross procedures may be
pursued for each character and the backcrossderived lines from each finally merged into a single line.
The backcross procedure is further utilized to transfer entire sets of chromosomes into a foreign cytoplasm to obtain cytoplasmic male sterility for the production of
hybrid seed as in
Page 174
Fig. 9.6.
Procedure for a backcross in which a dominant allele for disease resistance (R) is transferred from a
disease resistant cultivar to an adapted cultivar A. The resistant donor cultivar is crossed to the adapted
recurrent cultivar A, and the F1 backcrossed to cultivar A. The BC1 generation from this cross will be
segregating for disease resistance (Rr:rr). The Rr plants may be identified from the rr plants by
inoculating the seedling plants with the disease pathogen and observing whether plants exhibit the
resistant or the susceptible disease reactions. Only Rr (resistant) plants are backcrossed to A in the
second and succeeding backcross generations. After the final backcross, the heterozygous (Rr) plants
are selfed one generation to obtain homozygous (RR) and heterozygous (Rr) resistant plants. Progeny
tests of the resistant (RR and Rr) plants are grown to identify the homozygous (RR) from the
heterozygous (Rr) plants, so that lines pure for resistance may be established.
Page 175
corn, millet, onion, sorghum, wheat, and other crops. The species or cultivar with the foreign cytoplasm is the female and recurrent parent as cytoplasm is transferred
only through the egg. The donor of the chromosomes is crossed as the pollen parent until all donor chromosomes are recovered in the cytoplasm of the recurrent
parent. Normally, the original cross and four or five backcrosses are required.
Multiline Breeding
The traditional breeding procedures for selfpollinated crops were developed around the pureline concept. Pureline selection was utilized to isolate the superior plant
from genetically mixed populations, such as landraces, or from segregating populations following hybridization. Extreme uniformity among the plants within the selected
line was stressed. The uniformity often led to cultivars with a single gene for resistance to a particular pathogen being propagated over large geographic areas. If a new
race of the pathogen that was virulent on cultivars with the resistance gene arose, widespread disease damage would be caused. The rapidity with which new races of
disease pathogens arose sometimes limited the usefulness of cultivars with a particular resistance gene to no more than 5 to 10 years. This condition led to the
proposition that greater diversification in resistance genes would provide stronger genetic barriers and spread the risk from disease damage. One solution proposed
was the use of multiline cultivars.
As proposed to combat disease resistance, a multiline cultivar is a composite of genetically similar lines, except that each line possesses a different gene for
resistance to the disease pathogen (Fig. 9.7). Lines that are genetically identical, except for a single gene, are called isolines. The procedure for producing a multiline
cultivar is to develop isolines of a desirable cultivar, each with a different gene for resistance to a particular disease pathogen. Each gene should contribute resistance to
a different physiologic race, or group of races, of the disease pathogen. The backcrossderived isolines are then composited to form the multiline cultivar. As changes
occur in the prevalent races of the disease pathogen, isolines with new genes for resistance may be developed. Because the multilines are reconstituted each year, the
new genes for resistance may be introduced by changing the mix of the isolines.
Fig. 9.7.
Procedure for production of a disease resistant, multiline cultivar. Genes for
rust resistance, R1 to R5, are backcrossed from donor cultivars to a common
disease susceptible, recurrent cultivar A. Isolines are generated that differ
only in a gene for disease resistance and composited to synthesize the multiline
cultivar. The isolines are maintained so that the multiline can be resynthesized
as needed. Five crosses (original and four backcrosses) were made to the
recurrent cultivar.
Page 176
The theory of the multiline is to produce a population uniform for height, maturity, and other features, yet mixed in genes for resistance to pathologic races of a virulent
disease. The merits of the multiline cultivar are based on the assumption that it will provide partial protection to a broad spectrum of races of a diseaseproducing
pathogen and provide a buffering effect against rapid disease development should a new race of the disease pathogen arise. These assumptions have not yet been fully
tested.
The multiline cultivar concept as outlined has distinct disadvantages:
• a multiline has limited utility except in high risk areas where severe disease damage occurs regularly from a highly specialized disease pathogen,
• there is no genetic improvement in yield or other characteristics, except that provided by the disease resistance, as long as the same cultivar is used as the recurrent
parent for isolines development, (unless the recurrent parent is improved the multiline may soon become obsolete),
• the labor required to produce and maintain the isolines reduces the resources for improvement in cultivar characteristics other than disease resistance, and
• the release of a multiline cultivar is delayed until all of the isolines are produced and increased.
The term ''multiline" is sometimes loosely applied to mixtures of genetically diverse lines combined in various ways to buffer against environmental stresses. More
accurately, these mixed populations should be called composites.
Variety Blend
A variety blend is a composite cultivar produced by mixing seed of two or more cultivars, the suggestion being that a blend of genotypes will yield consistently higher
than the average of the pure component genotypes, due to the buffering effect against genotype × environment interactions, and will be more stable over locations and
years than a pureline cultivar. The advantage of the latter tends to diminish as the number of cultivars in the blend is increased. A variety blend will be less uniform in
appearance than a pureline cultivar. In making a variety blend, cultivars should not be mixed that will adversely affect uniformity in maturity, or features that will
reduce the quality of the product. Variety blends need to be reconstituted at regular intervals to maintain stable performance.
NonTraditional Breeding Procedures
Traditionally, breeding procedures for improving selfpollinated crops involved crosses between two parents, followed by several segregating generations; yieldtesting
began only after a high degree of purity was reached. When the cycle was completed, superior segregates were crossed to start a new breeding cycle. Opportunities
for gene recombination were limited by the narrow, twoparent gene pool, and opportunities to breakup linkage blocks were restricted by a 10 to 15 year breeding
cycle. To overcome these deficiencies, breeders sought ways to increase the contributions into the gene pool. This has led to such nontraditional breeding practices as
making large numbers of crosses, early generation yield testing and rapid elimination of lowyielding lines, and systems to increase potential gene recombination.
Page 177
Large Numbers of Crosses
In a hybridization breeding program to improve simply inherited characteristics like seed color or disease resistance, parents with the desired genes may be readily
identified. Choosing parents for crosses to improve characters with complex inheritance like yield, is less precise because the complementary genes contributing to
yield from a set of parents cannot be identified by examining the parents. Only after the cross has been made and the range of segregation examined can the success of
the cross be determined. This has led breeders to make large numbers of crosses to improve the odds that hybrid populations with superior transgressive segregates
will have been made. It is also necessary to grow large segregating populations to improve the odds that the plant with the superior genotype will be included. Because
resources are never unlimited, the breeder usually must choose between growing moderate populations of a large number of crosses or large populations of a
moderate number of crosses. The former is generally favored because it involves use of larger numbers of parent cultivars and increases the odds that cross
combinations with greater potential will have been made.
Early Testing
After a cross is made, the progenies from each cross must be evaluated to find the superior segregates, a task that greatly exceeds the labor of making the crosses.
One procedure to reduce growth in the testing program is early testing, i.e., the evaluation of crosses, or selections from crosses, in yield trials at an early generation.
With early testing, inferior segregates, or entire crosses, may be quickly eliminated and only the superior segregates carried to the advanced generations where
homozygosity and uniformity are attained. Selection and yield testing, concurrently, in the early generations reduces the population size rapidly so that relatively few
strains will be remaining by the F5 or F6 generation when a practical state of homozygosity is reached. Superior yielding lines identified in an early generation may
immediately be used as parents in new crosses, even though homozygosity has not been fully attained, thus reducing the generations required to complete the crossing
cycle and enabling the breeder to make more rapid progress.
Population Improvement
RECUR RENT SELECTION. Recurrent selection is a population improvement procedure designed to increase the frequency of desirable alleles for a
particular quantitative character by frequent intermatings among superior genotypes within the population. Ideally, superior genotypes are isolated after each
cycle of mating and intercrossed to produce the next generation. Considerable success has been achieved with population improvement procedures in crosspollinated
species, where random mating among plants occurs by natural means. Application of population improvement procedures to selfpollinated crops has shown similar
success, but is difficult to employ due to the labor involved in making the larger number of hand pollinations required to intermate the selected genotypes.
MULTIPLE CROSS. The multiple cross, also called convergent cross, is produced by crossing pairs of parents, and then crossing pairs of F1s until all parents
enter into a common progeny according to the following scheme:
Page 178
With this system, many potential genetic combinations exist, because every seed produced after the initial cross is potentially a new hybrid. A disadvantage is that
many undesirable combinations are also brought together. Exceedingly large numbers of hybrid seeds must be obtained in the second and later crosses if the maximum
number of genotypes is to be represented in the progenies.
MALESTERILEFACILITATED HYBRIDIZATION. Malesterile genes have been used in barley and other selfpollinated crops to facilitate crossing by
eliminating the need for emasculation. A common procedure is to incorporate a malesterile gene into a few standard cultivars by backcrossing. The malesterile
isolines produced are each pollinated from a group of cultivars and the crossed seeds bulked to plant a composite cross. With uncontrolled pollination, malesterile
flowers will be pollinated from malefertile plants, thus facilitating gene recombination within the population. To obtain maximum gene recombination, seeds are
harvested only from malesterile plants. During the segregating generations, malesterile plants may be handpollinated from selected malefertile plants, and the seeds
from these crosses used to plant the next generation.
Plant Breeding: A Numbers Game?
To generate the full range of potential genotypes from a cross it is necessary to grow large segregating populations. This concept, combined with the concept that large
numbers of crosses among heterozygous genotypes are needed to break linkages, has often led to the popular notion that plant breeding is a numbers game; that
chances for success in developing superior cultivars are enhanced by a proliferation in the number of crosses made, by increasing the size of the segregating
generations, and by growing a myriad of selections from each cross. The rarity with which really superior genotypes occur provides considerable validity to these
concepts.
The danger with the numbers approach is that the breeder may extend crossing and testing activities beyond the resources available needed to evaluate the materials
adequately. There appears to be little value in generating the rare genotype unless it can be identified efficiently in the breeding materials being grown. The physical size
of the breeding nursery, particularly for yield testing, may be extended by mechanization and standardization of procedures, and by computerization of data collection
and processing. But the visual evaluation of the breeding lines, and the final evaluation of the data generated, requires personal judgements, which cannot be wholly
mechanized or computerized and are limited in number by the breeder's time. Thus, the breeder needs to compromise between the number of recombinations desired
and the number that can be evaluated efficiently and accurately within the resources of the program. While the breeder should select rigorously for the desired
genotype and ruthlessly discard those
Page 179
that are inferior, it is important that there be accurate and dependable information on which to base these judgments.
In traditional breeding procedures, extreme purity was attained before yield testing was initiated. With increased numbers of crosses and early testing, yield testing in
the early generations is given more importance in the decision making. Yield testing provides a powerful tool in selection for resistance to environmental stress, or
disease, provided that those production hazards are present with sufficient intensity to affect yield each season. On the other hand, too much reliance on yield data,
without full consideration of other breeding objectives, often leads to proliferation of undesirable plant characteristics, such as increased height, late maturity, increased
disease susceptibility, or inferior quality.
How Breeding Procedures for SelfPollinated Crops Are Utilized
In this chapter the breeding procedures for selfpollinated crops are discussed. Earlier in this chapter it was stated that the superiority of a new cultivar is determined
by the combination of genes that it contains, not by the procedure by which the cultivar was developed. In developing a breeding program for a specific crop, the
breeder is faced with many decisions that transcends his decisions on breeding procedures. What are the objectives of the breeding program? How can cultivars
already available be improved? What germplasm resources are available to make these improvements? What is the most efficient breeding procedure for
accomplishing these objectives?
Wheat, rice, and soybean have been chosen as examples of how breeding programs are developed in selfpollinated crops. These are presented in chapters 14, 15,
and 16, respectively.
Study Questions
1. How does the pedigreeselection breeding procedure differ from the bulkpopulation breeding procedure?
2. In the backcross breeding procedure, does it make a difference whether the donor parent is male or female? Why?
3. How does a multiline cultivar differ from a blend?
4. What are the advantages and disadvantages of early generation testing?
5. Which breeding procedures discussed in this chapter could be used to improve crosspollinated crops?
Further Reading
Borlaug, N.E. 1959. The use of multilineal or composite varieties to control airborne epidemic diseases of selfpollinated crop plants. Proc. 1st Int. Wheat Genet.
Symp., Winnipeg, Canada, 1958. Univ. of Manitoba, Winnipeg.
Borojevic, S. 1990. Principles and methods of plant breeding. Elsevier, Amsterdam.
Briggs, F.N., and R.W. Allard. 1953. The current status of the backcross method of plant breeding. Agron. J. 45:13138.
Page 180
Browning, J.A., and K.J. Frey. 1969. Multiline cultivars as a means of disease control. Annu. Rev. Phytopathol. 7:35582.
Busch, R.H., and D.D. Stuthman. 1990. Selfpollinated crop breeding; concepts and successes. p. 2137. In J.P. Gustafson (ed.) Gene manipulation in plant
improvement II. Plenum Press, New York.
Fehr, W.R. 1987. Principles of cultivar development. Vol. 1. Macmillan Publishing Company, New York.
Jensen, N.F. 1952. Intravarietal diversification in oat breeding. Agron. J. 44:3034.
McFerson, J.K., and K.J. Frey. 1991. Recurrent selection for protein yield in oat. Crop Sci. 31:18.
McProud, W.L. 1979. Repetitive cycling and simple recurrent selection in traditional barley breeding programs. Euphytica 28:47380.
Olsson, G. (ed.). 1986. Svalöf 18861986, research and results in plant breeding. Ltsförlag, Stockholm.
Snape, J.W. 1982. The use of doubled haploids in plant breeding. p. 5258. In C. Broertjes (ed). Induced variability in plant breeding. Centre for Agricultural
Publication and Documentation, Wageningen, Netherlands.
Weiss, M.G. 1972. Cultivar versus variety. Crop Sci. 12:551.
Page 181
10. Breeding CrossPollinated and Clonally Propagated Crops
Breeding procedures in crop plants are designed to exploit the reproductive structure of the particular species. Thus, the breeding procedures used with cross
pollinated crops will differ from those used with selfpollinated crops. Furthermore, procedures may differ with different species of crosspollinated crop plants
because the species differ in the structure of the reproductive system. In this chapter models of breeding systems are presented that can be adapted to the breeding of
crosspollinated species. Modifications of the breeding systems that may be necessary to adapt the procedures for particular species, such as corn, potato, sugarcane,
or forage crops, are discussed in chapters dealing with those specific crops. Most species of commonly cultivated crosspollinated crops are seed propagated. Other
species, potato and sugarcane for example, that reproduce sexually in their native habitat with normal crosspollination are propagated asexually as clones when
cultivated. A breeding procedure appropriate for asexually propagated species such as sugarcane is also described.
Genetic Structure of CrossPollinated Crops
The genetic structure of a population is determined by its gene pool and the opportunity for genetic recombination. It is in these characteristics that crosspollinated
populations differ from selfpollinated populations.
Breeding CrossPollinated Versus SelfPollinated Crops
In the breeding of selfpollinated crops, the homozygous nature of the individual plant is exploited. A population of selfpollinated plants may be composed of
either a single,
Page 182
homozygous genotype, or a mixture of homozygous genotypes. Plant selection is effective only within populations that are mixtures of genotypes. Mass selection for a
particular phenotype in a mixed population of a selfpollinated crop will reduce genetic variability and increase the frequency of genes affecting the characters being
selected. Pureline selection in a mixed population isolates single homozygous plants, or plants largely homozygous, which can be increased and reproduced as a
breeding line or as a cultivar. With selection focused on plants in selfpollinated crops, characters with qualitative inheritance tend to receive major attention.
After superior lines are identified, the breeder resorts to artificial hybridization to promote gene recombination and produce hybrid populations in which selection may
again be practiced.
In the breeding of crosspollinated species, the heterozygous nature of the individual plant is exploited. In a population of a crosspollinated species, open
pollinated corn, red clover, or perennial ryegrass are typical examples, each plant has both homozygous and heterozygous loci, but it is the heterozygous loci that give
this group of plants their characteristic genetic structure. As a consequence of natural crosspollination, the genes are reshuffled each generation and regrouped into
new genetic combinations. While theoretically following the HardyWeinberg law of genetic equilibrium, with respect to any particular locus, factors upsetting genetic
equilibrium—nonrandom mating, mutation, migration, and natural selection—are also operating to restrict the outcome. With an almost limitless number of gene
combinations possible within the gene pool, almost never would two plants be found with identical genotypes. Under natural environmental influences, crosspollinated
populations are relatively fluid, in which genes favoring adaptation and increased seed production tend to increase at the expense of genes unfavorable for adaptation
or fitness to reproduce. In a breeding population, the shift toward more adapted genotypes may be accelerated by selection, and by environmental stresses to which
the breeding population is subjected.
In crosspollinated crops the focus of the breeder is on populations instead of individual plants, and more emphasis is given to quantitative inheritance in
breeding systems than in selfpollinated crops. Due to the extensive heterozygosity in crosspollinated crops, there is an abundance of phenotypic variation; hence,
cultivars of crosspollinated crops are less uniform than cultivars in selfpollinated crops. Genetic variability for qualitatively inherited characters may be drastically
reduced by rigid selection, but genetic variability in quantitatively expressed characters continues to be present, due to inability of the breeder to select accurately for
individual gene effects and to the influence of the genotype × environment interactions.
Progeny Test versus Combining Ability Test
An important difference between breeding selfpollinated and crosspollinated crops is found in the way the breeder evaluates breeding materials. In a selfpollinated
crop, in which individual plants tend to be homozygous, the genotype is reproduced in the progeny rather precisely and may be evaluated by progeny tests. In a cross
pollinated crop, individual plants are heterozygous and field grown plants will largely be pollinated by pollen from other heterozygous plants growing in the vicinity.
Under these conditions, or even if selffertilized, the genotype of a heterozygous plant is not faithfully reproduced in its progeny. Thus growing a progeny test of an
openpollinated plant does not provide information comparable to that obtained from growing a progeny test of a selfpollinated plant. What is obtained is an
evaluation of the progeny performance of a random selection of gametes from a mother plant combined with a random selection of gametes from pollen plants of
unknown origin. A more suitable test would be provided if the plant had been pollinated with a heterogeneous collection of pollen (gametes) of known origin.
Performance could then be compared among progenies
Page 183
of plants pollinated with the same source of pollen. A more precise comparison could be made by pollinating the plants with pollen from an inbred (homozygous) line.
Each pollen grain would then be identical and the progeny performance test would measure the result of combining the assortment of gametes from representative
mother plants with gametes of identical genotype from the tester line. A test comparing progeny performance of plants or strains pollinated with a known tester line is
called a testcross and evaluates the combining ability of the mother plants or strains with the common tester line. The average or overall performance of a plant or
genetic strain in a series of crosses with different tester lines is a measure of its general combining ability, whereas the performance of a plant or genetic strain in a
specific combination in comparison with the performance of other cross combinations is a measure of its specific combining ability. Combining ability tests are used
to identify desirable combinations of inbred lines to cross in the breeding of hybrid cultivars, or to identify desirable clones to include in a synthetic cultivar of a forage
crop.
Plant Features Promoting CrossPollination
Crosspollination may be promoted by morphological features or physiological characteristics of the plant. The more important of these are:
MONOECY. Monoecy is the separation of pistillate and staminate flowers on the same plant (see Fig. 2.4). Plants with monoecious flowers are normally cross
pollinated but some selfpollination may occur. Corn is an elegant example of an important crop with monoecious flowers. Because the flowers are borne on different
parts of the corn plant and a selfincompatibility system is not present, either self or crosspollinations are easily made.
DIOECY. Dioecy is the production of pistillate and staminate flowers on different plants (see Fig. 2.5) and promotes crosspollination. Dioecy is found in various
species of field crops (hops, hemp, and buffalograss), vegetable crops (asparagus, spinach), and fruit crops (date palm, papaya). With true dioecy selfpollination is
prohibited. However, because some normally dioecious plants produce monoecious flowers in addition, selfpollination on these plants is possible, unless restricted by
selfincompatibility systems.
SELFINCOMPATIBILITY. Selfincompatibility is the failure to become fertilized and set seed following selfpollination (see Fig. 7.1). Selfincompatible species
are normally crosspollinated due to the physiological hindrance of normal fertilization following selfpollination.
MALE OR FEMALE STERILITY. Male sterility, either genetic or cytoplasmic, promotes outcrossing because malesterile plants do not produce viable pollen
(see Figs. 7.2, 7.4, 7.6). With female sterility, the ovule does not function normally and seed production is inhibited. Female sterility is less common in plants than male
sterility.
FLORAL DEVICES. Various other floral devices promote crosspollination, such as failure of male and female sexual organs to emerge and mature at the same time
(see Fig. 7.7), or the need for insect manipulation of flowers to effect pollination.
Inbreeding
Selfpollination, or inbreeding, in crosspollinated crops leads to a decline in vigor and productiveness, a phenomenon observed by many early plant hybridizers who
failed to grasp
Page 184
its significance. Dr. G.H. Shull reported the deteriorating effect of inbreeding openpollinated corn (maize) in a report to the American Breeders Association in 1908,
and suggested that the breeder should strive to maintain superior hybrid combinations. This concept led to the development of procedures for breeding hybrid cultivars
in corn and other crops. In forage crops, inbreeding may lead to a reduction in fertility and seed production.
Breeding SeedPropagated CrossPollinated Crops
As in selfpollinated crops, the initial step in a breeding program is to assemble collections of germplasm of the appropriate species from different sources, which we
will refer to here as source populations. Each source population is a collection of heterozygous plants representative of the population from which it originated. The
source and extent of the germplasm collections will be influenced by the particular species with which the breeder is working and the objectives of the breeding
program. The germplasm collections may originate from productive commercial cultivars or hybrids, improved breeding populations, introduced germplasm from gene
banks or plant introduction centers, or local collections if the species is native to the area or has been grown continuously in the area for a long period of time. The
source populations are maintained either as seed or vegetatively propagated materials according to the species involved.
Breeding procedures in crosspollinated crops are based largely on population improvement principles, i.e., increasing the frequency of genes in the
population for the desired breeding objectives. It is imperative for genes that contribute to the enhancement of the desired objectives to be present in the source
population. Subjecting the source population and selections from it to environmental stresses, according to the objectives of the breeding program, will assist in
identifying the superior genotypes and their progenies. Although single plant selection is practiced extensively in the different breeding systems, the cultivar is not
established from a single plant in seedpropagated species because segregation and crosspollination make it impossible to maintain the parent genotype in succeeding
generations, and a wider range of genetic diversity than is found in a single plant is needed to maintain a vigorous and productive population. The selection of plants
from the source population from visual evaluation is referred to as phenotypic plant selection. Except in the mass selection procedure, initial selection is followed by
some form of progeny evaluation, either by growing a progeny test, or by crossing to a common tester genotype or a combination of testers and growing the
testcross progeny. Corn, an annual species, is used in the examples of the breeding procedures described here; with biennial or perennial species, the length of the
breeding cycles will require a larger number of generations than illustrated here with corn.
The RecurrentSelection Principle
Recurrent selection is any breeding system designed to increase the frequency of desired alleles for particular quantitatively inherited characters by
repeated cycles of selection. A recurrentselection cycle involves:
• identification in a source population of genotypes superior for the specific quantitative character being improved, and
Page 185
• the subsequent intermating of the superior genotypes to produce new gene combinations with improved expression of the character.
Selection cycles may be repeated as long as superior genotypes are being generated.
Phenotypic recurrent selection is selection to improve a plant quantitative character based on visual observation or physical measurement of the character. Examples
are oil content in corn (Fig. 10.1), fiber strength in cotton, sugar content in sugarbeets, or seed size in wheatgrass. Phenotypic recurrent selection is an appropriate
breeding procedure in naturally crosspollinated species or species where artificial crosspollinations are made easily.
Genotypic recurrent selection is selection to improve a plant quantitative character based on progeny performance as measured by test crosses, or by other means,
and is utilized to improve complex characters such as combining ability in corn inbred lines.
A model for simple phenotypic recurrent selection is illustrated in Fig. 10.2 in which plants are visually selected from the source population and the progenies of the
selected plants are grown and intercrossed to obtain new gene combinations. The crossed seed is used to grow a new source population, which starts the next
selection cycle. Selection based on phenotype will be effective insofar as the phenotype accurately identifies the superior genotype for the character under
consideration. Phenotypic recurrent selection is most effective for characters with low genotype x environment interaction, such as height of ears (in corn), seed size, or
disease resistance. For quantitative characters that cannot be selected accurately from the phenotype, breeding procedures have been devised based on progeny or
testcross performance that utilize the recurrent selection principle.
Fig. 10.1.
Comparison of oil percentages in a population of
corn after one and two cycles of recurrent
selection (Agron. J. 44:32931).
Mass Selection
Mass selection in crosspollinated crops is a selection procedure in which:
• individual plants are chosen visually for their desirable traits, and
• the seeds harvested from the selected plants are bulked to grow the following generation without any form of progeny evaluation (Fig. 10.3).
Repeating mass selection utilizes the recurrent selection principle. An example of the procedure follows:
First season: Select 50 to 100 plants with desired features from the source population and harvest openpollinated seed from each.
Second season: Plant a mixture of the seed harvested in the previous year. From this population, again harvest openpollinated seed from 50 to 180 plants selected
for desired features.
Page 186
Mass selection was one of the earliest procedures to be used in breeding corn, forages, sugarbeets, cotton, and other crosspollinated crops. In openpollinated corn,
mass selection was practiced by the farmer each season when he selected seed ears during harvest for planting the next crop. In forage crops, local strains stabilized
by natural selection may be harvested en masse and the seed utilized to start a new cultivar. Continuous mass selection for a specific character with high heritability that
can be evaluated visually, such as early flowering, will shift the gene frequency of the character in the direction of the selection. Curlytop resistance in sugarbeets was
increased by mass selection combined with recurrent selection (Fig. 10.4). Mass selection is simple to conduct; with its oneyear cycle, new cultivars can be
developed quickly and improvement is continuous. A weakness of mass selection is the lack of control over the pollen source and the gene contribution to the progeny
through the pollen gametes. For this reason, heritability estimates are reduced by onehalf. Selection for characters with low heritability is relatively ineffective, because
plants superior due to phenotype may not be distinguished from plants superior due to environmental influences.
Fig. 10.2.
Model for phenotypic recurrent selection. Note
that the mean of the populations has increased
following each selection cycle.
Fig. 10.3.
Mass selection. Individual plants are selected for desirable
traits and seed is composited to grow the next generation.
The composite may serve as the ''source population" for the
next selection cycle.
A procedure known as gridding may be used to reduce errors in selection caused by uneven environments. The land area on which the source population is grown is
divided into small plots or grids. Plants are evaluated within each grid and only one superior plant from each grid is harvested. This procedure gives equal
representation in the mass selection from all areas of the field regardless of field gradients in soil fertility or moisture supply. The grid system may be utilized when
making selections from the source populations in the halfsib, fullsib, and S1 procedures that follow.
Page 187
Fig. 10.4.
Progressive improvement in sugar beet for curlytopvirus resistance through a
combined mass/recurrentselection breeding procedure. Improvement was possible
because (1) genotypes with genes for resistance were present in the European source
population, (2) the severity of the curlytopvirus disease eliminated the susceptible
genotypes from the population, and (3) interpollination among the resistant genotypes
resulted in sugar beet plants with transgressive segregation for resistance. About 25%
of the plants in US 1 were resistant to the curlytopvirus, 40 to 50% of the plants
in US 33 were resistant, 75% of the plants in US 12 were resistant, and 85 to 90% of the
plants in US 22 were resistant.
HalfSib Selection with Progeny Test
Halfsib refers to a plant or family of plants with a common parent or pollen source. A halfsib selection procedure based on a progeny test differs from mass
selection because the new population is constituted by compositing halfsib lines selected from progeny performance rather than from phenotypic appearance.
Progenies of 25 to 50 plants are grown in replicated plots, so that the variance and mean performance may be evaluated. An example of the halfsib selection
procedure as used with corn follows (Fig. 10.5):
First season. Select 50 to 100 plants with desired features from an openpollinated source population, keeping the seed harvested from each plant separate. The
seed from each plant will constitute a different breeding line.
Second season. Using seeds harvested from openpollinated plants in the previous season, grow a progeny test of each line in an isolated area. Retain the remnant
seed.
Third season. The population is reconstituted by compositing equal quantities of either (a) seed harvested from the 5 to 10 superior progenies, or (b) remnant seed
from the 5 to 10 lines with superior progeny performance. Grow the composite in isolation with openpollination to obtain new gene combinations. In (a), onehalf of
the genes come from a random assortment of pollen from the lines in the progeny test, and in (b), onehalf of the genes come from a random assortment of pollen from
the original source nursery. As an alternative to growing the composite in isolation, plants may be hand pollinated from a composite of pollen. Seed harvested in the
third season may be:
Page 188
Fig. 10.5.
Halfsib selection based on progeny test performance.
• increased as a new openpollinated cultivar,
• planted as a source population to start a new selection cycle, or
• planted as a source population for isolation of new inbreds in a hybrid breeding program.
The halfsib selection procedure outlined, or variations of it, was utilized by some corn breeders in the early part of the century. Designated eartorow breeding,
individual ears were selected according to the breeder's fancy, the seed from each ear was planted in a separate row, and superior rows harvested for further selection
or seed increase. In corn, the eartorow procedure, as with mass selection, was effective in changing gene frequency for highly heritable characters like ear height that
could be selected visually, or like oil content of the kernel that could be measured. It was not effective in changing characteristics with low
Page 189
heritability such as yield. In those days, entries in the progeny test were not replicated and, without replication, variance due to genetic effects could not be separated
from variance due to environment. The halfsib procedure is applicable to crosspollinated crops, like corn or sugarbeets, where sufficient seed can be harvested from
a single plant to grow a yield trial, or to crosspollinated crops where selfpollination cannot be consummated due to selfincompatibility systems. Like mass selection,
it is based on maternal plant selection, without pollen control, in which case heritability estimates are reduced by onehalf.
HalfSib Selection with Testcross
In this procedure the selection of the halfsib lines to composite is based on testcross performance rather than progeny performance. An example of the procedure as
used with corn follows (Fig. 10.6):
First season. Prior to flowering, select 50 to 100 plants with desired plant characters from a source population: (a) pollinate a tester parent plant with pollen from
each of the selected plants and harvest crossed seed from the tester parent and openpollinated seed from the selected plants, keeping identity of each seed lot; or (b)
with pollen from each selected plant, pollinate a tester plant and selfpollinate the selected plant. Harvest crossed seed from tester parent plants and selfed seed from
selected plants, keeping identity of each seed lot.
Second season. Grow testcross progenies.
Third season. Reconstitute the population (a) by mixing equal quantities of openpollinated seed from 5 to 10 selected plants with superior testcross progeny
performance; or (b) by mixing equal quantities of selfed seed from 5 to 10 selected plants with superior testcross progenies. Grow the seed composite in an isolated
seed plot with openpollination to obtain new gene combinations.
The halfsib testcross procedure permits control over the testcross parents so that a more precise evaluation of the genotype of the selected plant is obtained than from
growing a progeny obtained by openpollination as in the previous procedure. If the tester is an inbred line, plants in each of the lines in the testcross progeny nursery
will have one parental gamete in common. Procedure (b) would be superior to procedure (a) because only genes from the plants with superior testcross progenies
enter into the gene pool of the composite, whereas in procedure (a) onehalf of the genes originate from a random selection of pollen from the source population.
The procedures outlined are applicable to corn and other crosspollinated crops in which sufficient seed can be produced by crossing to grow a replicated testcross
progeny trial. For procedure (b), selfpollination is necessary in addition, a requirement that could not be accommodated in selfincompatible species.
FullSib Selection
With fullsib selection, crosses are made between selected pairs of plants in the source population, with the crossed seed used for progeny tests and for reconstituting
the new population. An example of the procedure follows (Fig. 10.7):
First season. Cross 150 to 200 pairs of plants selected from the source population. Reciprocal crosses may be made to provide a larger quantity of crossed seed.
Second season. Grow a replicated progeny test with seed from each pair of crosses,
Page 190
Fig. 10.6.
Halfsib selection based on testcross progeny performance.
keeping the remnant crossed seed.
Third season. Reconstitute the source population by mixing equal quantities of remnant crossed seed from 15 to 20 paired crosses with superior progeny
performance, and grow in isolation with openpollination to obtain new gene combinations.
Fullsib selection measures the combining ability from mating specific pairs of plants and only those pairs with superior progeny performance enter into the composite.
The procedure is applicable to many crosspollinated species, including selfincompatible species.
Page 191
Fig. 10.7.
Fullsib selection based on progeny test performance of paired crosses.
Selection from S1 Progeny Test
S1 progeny tests may be utilized to evaluate selected plants from an openpollinated source nursery. S1 refers to the progeny following selfpollination of plants in an
openpollinated population, or in the F2 following a cross. The procedure follows (Fig. 10.8):
First season. Select 50 to 100 plants from a source nursery prior to flowering. Selfpollinate and harvest selfed seed from selected So plants.
Second season. Grow replicated S1 progeny trial, keeping remnant selfed (S0) seed.
Page 192
Fig. 10.8.
Selection based on S1 progeny performance.
Third season. Composite equal quantities of remnant seed from the So plants with superior progenies, and grow the seed composite in isolation to obtain new gene
combinations.
Selection is based on performance of S1 plant progenies. Each S1 progeny evaluated receives only the genes present in the parent So plant; no genes are introduced
into the line from openpollination or tester parents. The procedure is applicable to corn and other crosspollinated species in which a quantity of seed sufficient for a
replicated progeny trial and remnant seed for making the composite can be obtained by selfpollination. It would not be applicable to selfincompatible species.
Page 193
Reciprocal Recurrent Selection
Reciprocal recurrent selection is a procedure designed by corn breeders to improve two populations simultaneously for both general and specific combining ability.
Recurrent selection for general combining ability involves the use of a tester with a broad genetic base and identifies mainly additive genetic effects. Recurrent
selection for specific combining ability relies on a tester with a narrow genetic base and identifies both additive and nonadditive gene action. With reciprocal
recurrent selection, plants are selected in each of two populations, with the selected plants of one population being selfed and outcrossed as the tester to the selected
plants in the other population. After the testcross progenies are evaluated, remnant seed from the plants with superior testcross progenies are grown and intercrossed
to reconstitute the two populations. This completes the selection cycle.
Synthetic Cultivar
A synthetic cultivar is an advanced generation of a seed mixture of strains, clones, inbreds, or hybrids among them, propagated for a limited number of
generations by openpollination. The word "synthetic" implies a population of plants artificially produced by the breeder. The component strains, clones, or inbreds
are maintained, and the synthetic is reconstituted at regular intervals. It is incorrect to apply the term synthetic to populations originating from seed mixtures advanced
by openpollination without periodic reconstitution.
The synthetic procedure is widely used in breeding forage crops. In forage species, the halfsib, fullsib, or S1 selection procedures are rarely applicable because:
• the quantity of seed produced from a single plant is usually inadequate to grow a progeny test,
• selfincompatibility inhibits production of selfed seed in many forage species, and
• controlled crosspollinations are difficult to make in most forage species.
In addition to forage crops, synthetic cultivars may be developed in corn, sugarbeets, sunflower, and other crosspollinated species.
The design of the synthetic cultivar utilizes the partial exploitation of heterosis during a limited number of generations of seed increase. This feature has made the
synthetic cultivar popular in breeding forage species where conventional crossing procedures to obtain heterosis are not feasible. The procedure for developing a
synthetic cultivar has these essential characteristics:
• The synthetic cultivar is constituted from reproducible units of a crosspollinated crop (clones in forage species, inbreds in corn or sugarbeets).
• The plant materials entering into the synthetic cultivar are selected from performance in combining ability or progeny tests.
• The synthetic cultivar is constituted by random interpollination of the component units.
• The component units are maintained so that the synthetic may be reconstituted at regular intervals.
A model for the development of a synthetic cultivar in a forage species that embodies these characteristics is illustrated in Fig. 10.9. The procedure involves several
distinct populations.
Page 194
Fig. 10.9.
Procedure for development of a synthetic cultivar
of a forage crop, based on polycross progeny performance.
Source population. Plant selections are assembled from many sources to ensure a broad range of genetic variability. The original plant selections may come from old
established pastures or meadows, improved cultivars, introductions, bulked populations after several cycles of recurrent selection for a particular characteristic, or
other sources. The clones should be
Page 195
vigorous and productive so that they can be easily maintained and will produce vigorous and productive progenies.
Clonal line nursery. From the source nursery several hundred superior plants, are chosen and multiplied as clones for establishing a clonal line nursery (Fig. 10.10).
Each clonal line in the nursery will be comprised of 20 to 25 plants propagated vegetatively from the original plant. The clonal lines are evaluated for vigor, persistence,
and other superior characteristics, depending upon the species and the objectives of the breeding program. Exposure of the clones to adversities, such as severe
clipping, disease epidemics, cold, or drought, will aid in identifying clones with superior qualities. Finally, 25 to 50 of the superior clones will be selected for progeny
testing in a polycross nursery.
Making the polycross. Seed for growing a progeny performance test is obtained by making a polycross. The polycross is an isolated group of clonal lines
replicated in such a manner that each clone will be pollinated by a random sample of pollen from all of the other clones. The seed from each clone is
harvested separately with the identity of the clone maintained.
Polycross progeny test. The openpollinated seeds harvested from the clone in the polycross are planted in a progeny test for evaluation of yield and other
characters. From the polycross progeny test performance, 5 to 10 of the clones are chosen as components for the synthetic.
Syn 0 generation. The 5 to 10 clones chosen to be utilized in the synthetic are vegetatively propagated and randomly transplanted into an isolated seed field, or, in the
case of a legume, the clones may be transplanted into an insectproof cage and pollinated with bees to obtain seed. This constitutes the Syn 0 generation. Random
crosspollination among the Syn 0 clones fosters gene recombination.
Fig. 10.10.
Strains of weeping lovegrass. In the development of a synthetic cultivar in
forage crops, the characteristics of the individual breeding lines are evaluated
by growing the lines in a field nursery. The combining ability of the lines are
evaluated by growing the lines in a polycross yield nursery.
Page 196
Syn 1 generation. Openpollinated seed harvested from the Syn 0 generation is planted in isolation for seed increase. This constitutes the Syn 1 generation and may
be distributed as a synthetic cultivar if seed can be produced in sufficient quantity. Superior plant selections from the Syn 1 generation may be vegetatively propagated
to start a new source nursery.
Syn 2 generation. Openpollinated seed harvested from the Syn 1 generation is increased in isolation. This constitutes the Syn 2 generation.
The purpose of growing the Syn 2 generation is to increase the quantity of seed that will be available to the farmer. If sufficient seed to meet market demand can be
produced in the Syn 1 generation, it will be unnecessary to grow the Syn 2 generation. In most instances it is necesary to go to the Syn 3 or later generations to have
adequate seed for sale to farmers. The Syn 1 and Syn 2 generations are comparable to the F1 and F2 generations, respectively, in conventional hybridization. Each
generation the synthetic cultivar is advanced beyond the Syn 1 there will be successive reductions in vigor. The original clones are maintained so that the synthetic
cultivar can be reconstituted when the seed fields need to be renewed. The Syn 1 generation may be utilized as a source nursery from which to select clones that could
be used in breeding future synthetics, thus introducing the recurrentselection principle. In addition to polycross performance, as illustrated here, to evaluate clones, the
clones may be evaluated by S1 performance trials.
A synthetic cultivar in corn is produced in a similar manner to that illustrated with a forage crop except that the breeder will be working with inbreds instead of clones.
The corn inbreds to be used as component lines in the synthetic cultivar are chosen on the basis of combining ability tests, and crossed in all combinations to produce
the seed for growing the Syn 1 generation. The inbreds are maintained so that the synthetic can be reconstituted. Similar procedures may be utilized to produce
synthetic cultivars in other crosspollinated crops. In annual species of plants such as corn, the synthetic will need to be reconstituted each season, otherwise, the
population behaves as an openpollinated cultivar.
Choice of Source Germplasm and Test Environment
With recurrentselection procedures, the performance level of the source population and the test environment are important. A superior combination of alleles for the
character being improved will not be forthcoming if the genes are not in the source population. In choosing component lines for the source population, only those with
the highest expression of the desired characters should be included. If the lines have a diverse origin, there may be a greater possibility that they will contain different
alleles for the character. When evaluating the populations, they will need to be grown in an environment that fosters expression of the character to be improved, if the
superior genotype is to be identified.
Breeding Clonally Propagated Crops
A clone is a vegetatively propagated population of genetically identical plants. In asexually propagated species, the separate genotypes are propagated as clones.
Clonal propagation may be practiced with species that produce seeds poorly or that produce seeds only under special conditions. Some crops normally propagated
as clones are sugarcane (see Fig. 2.10), potato, sweet potato, cassava, sisal, taro, and some species of perennial grasses, such as bermudagrass. Asexually
propagated species have not normally been subjected to selfpollination and
Page 197
inbreeding, and individual plants are highly heterozygous, the heterozygosity being maintained through clonal propagation. With potatoes as an exception, most clonally
propagated species are perennials. Aneuploid or polyploid chromosome genomes are maintained with clonal propagation, resulting in clones with chromosome
numbers that differ from those recorded for the species. Breeding procedures for clonally propagated species may be grouped into:
• germplasm assembly and maintenance,
• clonal selection of natural or induced variants, and
• hybridization followed by selection and propagation of superior clones in the segregating population.
Germplasm Assembly and Maintenance
As with crops that reproduce sexually, the initial step in breeding asexually propagated species is to assemble a germplasm collection that is maintained as clones. The
germplasm assembly may include clones selected from local populations if the species is native to the locality, introduced clones from genebanks or other breeders,
commercially grown cultivars, or wild relatives introduced from their native habitat. The germplasm collection of clones constitutes the breeder's source nursery.
Clones from the source nursery may be propagated and grown directly as cultivars, or the clones may be used as parents in a hybridization program. The germplasm
collection is maintained as a collection of living plants in the field; this differs from maintaining a seed collection as in a sexually propagated species. Because vegetative
propagation maintains the genotypes without change, except for mutation, large numbers of clones may be grown in the breeding nursery without isolation. In most
countries it is mandatory that clones introduced from a foreign source first be grown in isolation to prevent the possible introduction of new species of insects or
disease pathogens along with the clone. The hazard may be reduced by introduction of seeds instead of clones, if the species produces viable seeds.
Clonal Selection
In a genetically mixed population of an asexually propagated species such as exists in nature, a superior clone may be isolated and propagated as a cultivar. In a mixed
population, progress through clonal selection is limited to the isolation of the best genotype present. Genetic variability may arise in a clone by mutation producing bud
sports, chimeras, or genetic mosaics. In species of ornamental plants, variants originating from natural or induced mutations are often utilized as the source of new
clones. A high mutation rate has been observed in genotypes of sugarcane maintained through tissue culture techniques, with the mutant plants then propagated as
clones.
Hybridization
Gene recombination occurs with sexual reproduction. In a crop species that is normally propagated asexually, sexual reproduction is necessary to create
genetic variability through gene recombination. By crossing clones with superior characters, source populations will be created that may be utilized for the
selection of new clones as in selfpollinated crops. A typical procedure for developing a cultivar from an asexually propagated species, such as sugarcane, follows
(Fig. 10. 11):
Page 198
Fig. 10.11.
Hybridization procedure for a clonally propagated species. The seedlings
grown in the first season are comparable to an F2 generation in a conventional
hybridization procedure. The genotype of each seedling plant is maintained by
vegetative propagation in the first and succeeding seasons.
Crossing generation: Cross Clone A × Clone B.
1st season: Grow 10,000 F1 seedling plants. Select 1000 vigorous plants and propagate vegetatively.
2nd and 3rd seasons: Grow 1000 clonal rows in 2nd season; select 100 superior clones. Grow 100 clones in 3rd season, preferably at two locations; select 10
superior clones.
4th to 7th seasons: Grow selected clones in replicated field trials at several locations in comparison with standard cultivars or advanced breeding lines.
8th to 10th seasons: Increase propagules of superior new clone and release as a new cultivar.
Number of seedling plants and clones to grow are suggestive only and will vary with the species and resources of the breeding project.
Page 199
Due to the openpollination, the parent clones will be heterozygous, segregation occurs in the F1 generation; each F1 plant is thus a potential source for a new clone
and a new cultivar. Clones propagated from F1 plants are heterozygous and the heterozygosity of the clone is maintained through asexual propagation. If the breeder
does not find a superior genotype in the F1 generation, the crosses are remade, or different crosses may be made. Selfpollination to produce an F2 is seldom practiced
because selfpollination leads to a reduction in vigor and fertility. If a superior F1 plant is identified in the hybrid progeny, it is propagated vegetatively to establish a
new clone which is evaluated in observation and replicated plot tests.
How Breeding Procedures Are Utilized
In this chapter the breeding procedures for crosspollinated and clonally propagated species were discussed. For discussion of how the procedures are utilized in
particular crops, the student is referred to Chapter 19 on Breeding Cotton and Chapter 20 on Breeding Forage Crops. For discussion of how the procedures are
utilized in breeding clonally propagated crops, the student is referred to Chapter 21, Breeding Potato, and Chapter 22, Breeding Sugarcane.
Study Questions
1. What is mass selection? How can the breeder modify mass selection to improve upon its effectiveness as a plant breeding method?
2. What is the difference between halfsib and fullsib recurrent selection?
3. Why is recurrent selection a more efficient breeding method in crosspollinated crops than selfpollinated?
4. How does the breeding of clonally propagated crops differ from the breeding of nonclonally propagated crops? Why is the first generation in a cross between
clonally propagated plants the same as the F2 in a regular hybridization program?
Further Reading
Abbott, A.J., and R. Atkin. 1988. Improving vegetatively propagated crops. Academic Press, Inc., San Diego, CA.
Cook, R.E. 1983. Clonal plant populations. Am. Sci. 71:24453.
Fehr, W.R. 1983. Principles of cultivar development, theory and technique. Vol.1. Macmillian Publishing Co., New York.
Frey, K.J. 1983. Plant population management and breeding. p. 5558. In D.R. Wood (ed.) Crop breeding. Am. Soc. Agron., Crop Sci. Soc. Am., and Soil Sci.
Soc. Am., Madison, WI.
Hayes, H.K., E.H. Rinke, and Y.S. Tsiang. 1944. The development of a synthetic variety of corn from inbred lines. J. Am. Soc. Agron. 36:9981000.
MorenoGonzález, J., and J.I. Cubero. 1993. Selection strategies and choice of breeding methods. p. 281313. In M.D. Hayward, N.O. Bosemark, and I.
Romagosa (eds.) Plant breeding, principles and prospects. Chapman & Hall, London.
Tysdal, H.M., and B.H. Crandall. 1948. The polycross progeny performance as an index of the combining ability of alfalfa clones. J. Am. Soc. Agron. 40:293306.
Vogel, K.P., and J.F. Pealersen. 1992. Breeding systems for crosspollinated perennial grasses. p. 25174. In J. Janick (ed.) Plant Breeding Reviews. Vol. 11. John
Wiley and Sons, New York.
Page 200
11. Breeding Hybrid Cultivars
Hybrid cultivars are the first generation offspring of a cross between inbred line parents with different genotypes. In the breeding of hybrid cultivars,
homozygous parent lines, generally called inbred lines, are first created by inbreeding (selffertilization) in a crossbreeding population. A hybrid cultivar differs from
a cultivar produced by hybridization as described for breeding selfpollinated crops in Chapter 9. With hybridization in selfpollinated crops, homozygous parent lines
are crossed and the segregating progenies are selfpollinated for several generations for the segregants to reach homozygosity, after which a pure line is selected and
grown as the new cultivar. With the hybrid cultivar, homozygous inbred lines (pure lines) are crossed and the heterozygous F1 generation is grown as the hybrid
cultivar.
The hybrid cultivar is produced in three steps:
• the development of inbred lines, normally by several generations of inbreeding in a natural or segregating population of a crosspollinated species,
• crossing pairs of unrelated inbred lines to produce a singlecross F1 hybrid cultivar with many heterozygous loci, and
• producing seed of the singlecross hybrid cultivar for distribution to the grower.
Because the inbred parent lines of the singlecross hybrid are homozygous at all loci, being comparable to a pure line in this respect, singlecross hybrid plants with
common parentage will be identical in genotype and uniform in appearance. Not only is the genotype of the superior singlecross plant reproduced in every plant in the
farmer's field, but the genotype is reproducible in succeeding years if the purity and identity of the inbred lines used in the cross are maintained.
The Origin of Hybrid Breeding
Hybrid breeding began in 1909 when George H. Shull proposed a method for producing hybrid cultivars of corn (maize). The previous year, Shull had reported that a
field of open
Page 201
pollinated (naturally crosspollinated) corn is composed of many complex hybrids which decline in vigor with inbreeding, and that the breeder should strive
to maintain the best hybrid combination (Fig. 11.1). From observations made while inbreeding and crossing in corn, Shull outlined a procedure for developing
inbred lines (pure lines) in corn, and crossing the inbred lines to produce singlecross hybrid cultivars. The proposal completely revolutionized corn breeding, and
breeding of hybrids has since been extended to other field and horticultural crops.
Edward M. East, working at the Connecticut Agricultural Experiment Station, also reported on the inbreeding of corn in 1909, but did not describe a clear procedure
for utilization of the inbred lines. Later, East and his students, Donald F. Jones and Herbert K. Hayes, made significant contributions to the development of procedures
for breeding hybrid corn. It appeared at first that the cost of producing the hybrid seed would limit the use of the system for breeding hybrids, because the seed was
being produced on weak and unproductive inbred plants. This problem was solved by Jones who, in 1918, proposed crossing pairs of inbred lines to produce single
cross hybrids, and then crossing two unrelated singlecross hybrids to produce a doublecross hybrid cultivar, according to the scheme outlined below:
The doublecross procedure made the production of hybrid seed corn economically feasible because the final hybrid seed was now produced in quantity on vigorous
and productive (female) singlecross hybrid plants, rather than weak and unproductive inbred plants. Pollen was also produced more abundantly by the vigorous
singlecross male parent.
Shull was not the first to observe the presence of hybrid vigor. Köelreuter, as early as
Fig. 11.1.
Openpollinated plant of corn and its progeny. (A) Plant of openpollinated corn.
(B) Ear from openpollinated plant: each kernel came from a separate fertilization,
the pollen originating from many different plants within the field. (C) Plants
grown from the seed of an openpollinated cultivar of corn. Each plant is
a different hybrid and varies in plant and ear characteristics, and in yield. In 1909,
Shull proposed that the objective in corn breeding should be to maintain the
genotype of the most vigorous plant, a suggestion that led to development
of hybrid corn.
Page 202
1763, had recorded the luxuriant growth of tobacco hybrids; Darwin, in 1877, concluded that the effect of crossfertilization in plants was generally beneficial,
whereas selffertilization was injurious; and in 1880, W.J. Beal in Michigan reported that variety hybrids of corn were more productive than openpollinated varieties.
Beal's procedure for making variety hybrids was to cross openpollinated varieties by detasseling rows of one variety which were then pollinated by a second variety
planted in adjacent rows. The procedure was remarkably similar to today's procedure for producing singlecross hybrid cultivars. But farmers in those days were not
ready for such radical innovations.
Before hybrid corn, it was the practice of corn breeders to breed for uniformity by repeated selection for particular plant, ear, and kernel characteristics. In
practice, the procedure was a mild form of inbreeding that led to a reduction in the heterozygosity of the breeding lines, and gradual loss in vigor and productivity. So,
for Shull to recommend growing a hybrid cultivar was revolutionary indeed. Years later (1952), Shull recalled just when he conceived the hybrid concept:
At the end of 1906, I had only the concept held by Holden, Shamel, East, and all other corn breeders who had experimented with the selfing of maize, that selfing had deleterious
effects, not that crossing has advantageous effects, other than the simple avoidance of the deleterious effects of selfing.
Quoting from his notebook on his 1907 experiments:
The obvious results were the same as in 1906, the selffertilized rows being invariably smaller and weaker than the corresponding crossfertilized. [But] a very different explanation
of the facts was forced upon me .... It may be assumed that correct field practice in breeding of corn must have as its object the maintenance of such hybrid combinations as
prove to be most vigorous and productive and give all desired qualities of ear and grain.
Shull's concept portended the procedures used today in breeding hybrid cultivars in corn and other crop species.
Inbreeding in CrossPollinated Crops
Inbreeding consists of any system of mating that leads to an increase in homozygosity. Inbreeding occurs when individuals are mated that are related by
ancestry. The most rapid approach to homozygosity in plants is through selffertilization; heterozygosity in a population of plants being reduced by onehalf with each
successive selffertilization(see Fig. 9.1). With selffertilization, heterozygous alleles (Aa) segregate into the genotype combinations, 1AA:2Aa:1aa; homozygous alleles
(AA and aa) continue to reproduce the same homozygous genotypes. As homozygosity increases in the breeding population, genotypefrequency changes, although
genefrequency remains unchanged (Chapter 4).
Among the naturally crosspollinated species, corn is a particularly favorable crop in which to study the effects of inbreeding because selfpollinations are easily made
and good seed set is normally obtained following selfpollination. In some crosspollinated crops, red clover for example, seeds are seldom obtained following self
pollination because of selfincompatibility. In plants where the sex organs are on separate plants, or if for other reasons selfpollination does not lead to normal
production of seeds, inbreeding may be accomplished through sib
Page 203
matings, the mating of close relatives. A halfsib mating is a mating between plants that have one parent or one pollen source in common (see Figs. 10.5, 10.6). A
fullsib mating is one between plants within the progeny of a single plant (see Fig. 10.7). Sib matings increase homozygosity, but complete homozygosity is reached
more slowly than through selffertilization. Ten generations of fullsib matings are normally required to reach the same level of homozygosity as three generations of
selffertilization (Fig. 11.2).
A major visible consequence of inbreeding in crosspollinated species is loss in size and vigor in progeny plants as heterozygosity decreases. The decrease in vigor is
largest following the first generation of inbreeding and levels off as homozygosity is approached (Fig. 11.3). The decline in vigor with inbreeding is known as
inbreeding depression and results from increases in the frequency of homozygous loci with deleterious effects. In plants with heterozygous loci, the recessive
deleterious allele is not expressed in the plant phenotype due to masking by the favorable, dominant allele. As homozygosity increases, many dominant alleles are lost
and the deleterious effects of the recessive alleles on the phenotype are expressed. In forage species such as alfalfa, or various species of grasses, fertility and seed
production are frequently suppressed following inbreeding. The decline in fertility often occurs so rapidly that the inbred progenies cannot be maintained by seed
propagation, or can be maintained only with great difficulty beyond the first or second selfed generation. With inbreeding in a crosspollinated species such as corn,
many individual plants in the progenies of the selfed plants exhibit serious faults, such as reduction in plant stature, weak stalks and a tendency to lodge, increased
susceptibility to disease pathogens, and a wide assortment of other deleterious plant characteristics (Fig. 11.4). In the development of inbred lines, the undesirable
plants are discarded, retaining only the most vigorous plants, which are selfpollinated again and the process repeated. The most important trait for an inbred line, its
contribution to yield in a hybrid combination, cannot be visually determined. This trait is expressed in the progeny when the inbred line is crossed with another inbred
and the hybrid progeny grown in a yield trial. In species of crop plants evolved through natural selfpollination, such as barley, beans, rice,
Fig. 11.2.
Comparison of homozygosity reached by selffertilization
( ).
Page 204
Fig. 11.3.
Reduction in vigor in corn with successive generations of inbreeding. S0 represents the
original selfed (inbred) plant and S1 to S6, successive selfed (inbred) generations.
soybean, tomato, or wheat, an adverse effect from inbreeding is not generally apparent and these crops can be maintained in a homozygous condition without apparent
loss of vigor.
Hybrid Vigor or Heterosis
Hybrid vigor is the increase in size, vigor, or productivity of a hybrid plant over the average or mean of its parents. The latter being referred to as the
midparent value. An alternative term, heterosis, was proposed by Shull to denote the stimulation in size and vigor in a hybrid as an expression of hybrid vigor. The
two terms, hybrid vigor and heterosis, are synonymous and may be used interchangeably. To be useful, the hybrid plant needs to exceed the best parent in yield and
productivity. Unless a hybrid is superior to its best parent line, it has no advantage for the breeder or the farmer. The effects of hybrid vigor in plants are manifested
most often by increased vegetative growth and yield of the harvested product; but hybrid vigor also may be reflected in cell size, plant height, leaf size, root
development, ear or head size, grain number, seed size, and other ways. Hybrid vigor is generally greatest following crosses among diverse genotypes of a cross
pollinated species, but also may be expressed following crosses among diverse genotypes in a selfpollinated species.
Explanations of Hybrid Vigor
Two theories are generally offered to explain the phenomenon of hybrid vigor. Neither appears to be wholly adequate. The most widely accepted explanation is based
on the assumption that hybrid vigor results from bringing together an assortment of favorable dominant
Page 205
Fig. 11.4.
Selfpollinated plant of corn and its progeny. (A) S0 plant from openpollinated cultivar.
(B) Ear from S0 plant. (C) S1 (first generation selfed) plants. Undesirable plants (2,3,5)
are discarded; desirable plants (1,4,6) are again selfpollinated. Selection and
selfpollination are continued for five to seven generations to obtain uniform inbred lines.
genes. According to this theory, alleles that contribute to vigor and growth are dominant, whereas the recessive alleles may be neutral, harmful, or deleterious to the
individual. If the dominant alleles contributed to the hybrid by one parent complement those contributed by the other parent, the F1 will then have a more favorable
combination than either parent. The theory works like this: for simplicity, let us assume that in corn the dominant genes ABCDE are favorable for high yield, that inbred
A has the genotype AABBccddEE (ABE dominant), and inbred B has the genotype aabbCCDDEE (CDE dominant). Genotypes of the inbreds A and B and the F1
hybrid are as follows:
In this simplified example, the F1 hybrid contains dominant genes at five loci (ABCDE) compared to three loci only in each of the inbred parent lines and would exhibit
more vigor than either of the parent inbred lines.
With the favorabledominantgene theory of explaining hybrid vigor, the question arose as to why favorable dominant genes could not be concentrated sufficiently in an
inbred line so that it would be as productive as a hybrid. In a crosspollinated population, the deleterious effects of many recessive genes are masked by the presence
of dominant alleles; with each selfpollination, approximately onehalf of the heterozygous loci become homozygous for recessive genes and contribute to the decline in
vigor observed in the inbred line. The theory assumes that in a crosspollinated crop like corn, there are too many deleterious recessive alleles to recover sufficient loci
with homozygous dominants for an inbred plant to exhibit the level of vigor of a hybrid plant. Much progress has been made in obtaining more favorable combinations
of the favorable dominant genes since the theory was proposed, especially in corn inbred lines. Inbred lines currently used in hybrid corn production are greatly
improved in
Page 206
vigor and productiveness over those used earlier. Since the early 1960s, farmers have been growing singlecross hybrids from seeds harvested from inbred lines,
making the double cross in hybrid seed corn production no longer necessary.
Another theory explains hybrid vigor on the basis that loci that are heterozygous contribute more to productivity than loci that are homozygous; the most vigorous
hybrid plant being the one with the greatest number of heterozygous loci. This theory is based on the supposition that there are contrasting alleles, for example, a1 and
a2, for a single locus. Each allele produces a different favorable effect in the plant. In a heterozygous (a1a2) plant, a combination of the effects is produced that is more
favorable to the plant than the effect produced by either of the alleles alone. The phenomenon of the heterozygote (a1a2) being superior to the homozygote, (a1a1) or
(a2a2), is termed overdominance and is based on interaction of alleles at the same locus. The theory of explaining hybrid vigor on the basis of heterozygous alleles
is defended by results from specific experiments, but neither the favorable dominant gene nor the overdominance theory satisfactorily explains all cases of hybrid
vigor. In general, the favorable dominant gene theory is favored by most experimental evidence.
Utilization in Cultivar Development
The principle of hybrid breeding is to mate homozygous parent genotypes that will combine producing superior F1 hybrid vigor, and to reproduce the superior F1
genotype in every plant in the hybrid population. Although hybrid plants are heterozygous at many loci, uniformity is attained, as in an inbred line, or a pureline cultivar
of a selfpollinated crop, or a clonally propagated line, because all hybrid plants have identical genotypes.
For hybrid cultivars of agronomic or horticultural seedpropagated crops to be accepted and used, the F1 hybrid seed must be produced in sufficient quantity to supply
grower demand at an affordable price. This requires a practical form of pollination control in the production of hybrid seed. In corn, a monoecious flowering plant,
pollination in seed production fields is controlled by removal of the tassel containing the male floral organs from the seedproducing plants, a process called
detasseling. The detasseled plants are pollinated from windblown pollen from adjacent malefertile pollinator rows. This procedure is not available in species in
which the flower contains both male and female floral organs. In some species, like cucumber, pepper, or tomato, where large numbers of seeds can be produced
from pollination of a single floret, hand emasculation and pollination are used to produce hybrid seed. But in field crops, hand pollinations are too expensive to
produce the needed quantities of hybrid seed. Cytoplasmic male sterility and fertilityrestorer genes are utilized, or have been proposed for utilization, for
controlling pollination in crop species in which usable forms of cytoplasmic sterility and restorer genes are available. Additionally, chemical hybridization agents that
prevent pollen formation offer a potential alternative to the use of cytoplasmic male sterility.
Breeding SingleCross Hybrid Cultivars
The singlecross hybrid cultivar is the product of a cross between two inbred lines or pureline cultivars. The production of singlecross hybrids began with the
research of Shull in corn in 1909 and was an intrinsic part of the doublecross corn breeding procedure because two single crosses were utilized in the production of
each double cross. From this background, the singlecross hybrid in corn became the model for breeding singlecross hybrid cultivars in other
Page 207
field and horticultural crops—onion, pearl millet, sorghum, sunflower, tomato, wheat—as well as in corn.
Inbred Line Development
An inbred line is a homozygous breeding line developed and maintained by selfpollination. When developing inbred lines in a hybrid breeding program in a
crosspollinated species such as corn, breeders normally start by selfpollinating heterozygous plants. The heterozygous plants may have been selected by the breeder
from (a) a natural population of a crosspollinated species, or a crosspollinated population improved through a recurrent selection procedure, or (b) the second
generation (F2) progeny from a cross between homozygous parent lines. In breeding hybrid corn the inbred lines were originally developed by selfing selected
heterozygous plants from fields with uncontrolled pollination (also referred to as openpollination) as in (a) above. In a mature hybrid corn breeding program, inbred
lines of corn are more generally developed from the hybrid progeny created by crossing two elite inbred lines as in (b). The latter is identical to the hybridization
procedure described for selfpollinated crops in Chapter 9. In (a), the original heterozygous selfed plant is normally referred to as the S0 plant, and the progeny
obtained from selfing this plant as the S1 (firstgeneration selfed) progeny. The secondgeneration selfed progeny are called the S2, and so on. If the heterozygous
plant originates from a cross between homozygous inbred lines, the F1, F2, etc., designation would be used as in selffertilized species.
The purpose of inbreeding is to reduce the offspring of a heterozygous plant into an array of dissimilar, homozygous, inbred lines. Striking differences are observed
between lines with successive generations of inbreeding; within lines, plants become more alike and the individual inbred lines become more distinguishable from each
other. With successive generations of inbreeding, homozygosity and uniformity are increased within the progeny lines. The variance within lines is reduced while the
variance between lines is increased. In a crosspollinated crop like corn, five to seven generations of selffertilization and pedigree selection are necessary to obtain
inbred lines that are uniform in plant and seed characteristics and that will remain uniform under continued selffertilization. This procedure is identical to the pedigree
selection procedure utilized in the breeding of selfpollinated crops, where a cross between selected parents is followed with several generations of selfpollination and
selection to generate uniform, truebreeding progeny lines (Fig. 9.2). During inbreeding, many undesirable recessive alleles at heterozygous loci will be replaced by
dominant alleles and the recessive alleles eliminated from the progeny; at other loci, recessive genes will become homozygous contributing to an overall loss of vigor in
the inbred. Greatest vigor is lost during the early generations of inbreeding, the loss in vigor declining as the inbred lines gradually approach homozygosity. After no
further loss of vigor is experienced, the genotype of the inbred may be maintained by selfpollination, unless mutations or outcrossing occur.
Selection during the early generations of inbred line development is based largely on visual observations of the inbred plants for characteristics that will affect the
suitability of the inbred to be utilized in commercial hybrid production. Are the plants strong and vigorous? Do they stand without lodging? Do they produce seeds in
sufficient abundance and quality to assure maintenance of the inbred line? Are the plants free from insects and disease? How do they yield?
During the period of inbreeding and selection, it is desirable to subject the partially inbred lines to adversities, such as high plant density, unfavorable high or low
temperatures, drought, lodging, disease, or insect pests, as appropriate for the particular species being inbred and the
Page 208
objectives of the breeding program. By such tests it is possible to identify the inbred lines that will give the most reliable performance in a wide array of environments.
It is also important in inbred line development to find new inbred lines that will contribute to greater productivity when crossed with other inbred lines.
Combining Inbreds into Single Crosses
The utility of the inbred line is determined by its genetic contribution to the hybrid progeny when crossed with another inbred line, not in its production potential per se.
However, it needs sufficient production potential so that it can be economically maintained and utilized as a parent line in the production of hybrid seed. Vigor and
productiveness that were lost during inbreeding are recovered in the hybrid when the inbred lines are crossed. From experience it has been learned that some inbred
lines will combine with a large number of other inbreds to produce highyielding hybrid progenies; other inbred lines will combine satisfactorily with few or no inbred
lines. The ability of the inbred line to transmit desirable performance to the hybrid progeny is referred to as combining ability.
GENERAL COMBINING ABILITY. General combining ability (gca) of an inbred line is the average contribution that the inbred makes to hybrid
performance in a series of hybrid combinations in comparison to the contribution of other inbred lines to hybrid performance in the same series of hybrid
combinations. It is not possible from visual observation to predict the contribution of an inbred line to hybrid performance. The gca of an inbred line is evaluated by
crossing it with other inbred lines and comparing the overall performance of the singlecross progenies. General combining ability evaluates the additive portion of the
genetic effects. If inbred lines A, B, C, D, and E are crossed in all possible combinations (diallel mating) and the singlecross hybrids are grown in a yield trial, the
inbred whose single crosses have the highest average yield would have the greatest gca. If that inbred is A, the implication is that A will contribute to high yield in a
wider array of crosses than inbred lines B, C, D, or E. With large numbers of inbred lines, it is not always feasible to make all possible diallel matings and grow the
hybrid progenies in performance trials. The number of possible singlecross combinations that can be made from n inbred lines is equal to n(n 1)/2. With 10 inbreds,
the possible number of singlecross combinations is 45; with 100 inbreds, the possible number of singlecross combinations is 4950, an impossible number to produce
or grow in a performance trial. These burdensome numbers made it clear to the early breeders of hybrid corn that a simple and efficient system of screening inbred
lines for combining ability was needed before pairing the inbred lines in singlecross yield trials.
From testcross experiments (often called topcrosses in corn), it was demonstrated that yields of corn inbred lines pollinated with a mixture of pollen, such as from an
openpollinated cultivar or from a doublecross or singlecross hybrid, were highly correlated with the average performance of the inbred line in a wide array of single
cross combinations. This led to the use of testcrosses for preliminary screening for gca of large numbers of newly developed inbred lines. As new inbred lines were
generated, they were first pollinated with a heterogeneous genotype proven to be an efficient tester, and the testcross progeny evaluated in yield trials. Only inbred
lines with superior gca were retained for testing in singlecross combinations.
SPECIFIC COMBINING ABILITY. Specific combining ability (sca) is the contribution of an inbred line to hybrid performance in a cross with a specified
inbred line, in relation to its contributions in crosses with an array of specified inbred lines. Specific combining ability
Page 209
evaluates nonadditive gene action and is utilized to identify the inbred x inbred cross combination with superior performance. The inbred lines identified as having
superior gca are crossed in all possible pairs (diallel crossing) to create single crosses, which are then evaluated in yield trials for sca. For example, if all possible
single crosses among inbreds A, B, C, D, and E are made, and the combination A × E produces the highest singlecross yield performance, the A × E cross
combination would have superior sca. Whether two particular inbreds combine to produce a highyielding single cross depends upon the extent that the favorable
genes for yield from the two parent inbreds complement each other. Experience has shown that inbreds derived from unrelated populations will combine to produce
highyielding single crosses more frequently than inbreds derived from related parent material.
BROAD VS. NARROWBASED TESTERS. In early hybrid corn breeding programs, testers with a broad genetic base, such as openpollinated cultivars, were
used to evaluate gca of corn inbred lines. As the hybrid breeding programs in particular species, such as corn, became more advanced, changes were made in the way
in which the breeder goes about the task of screening inbred lines for combining ability and fitting the lines into hybrid combinations. The gca of new corn inbreds is
now evaluated more frequently by crossing with narrowbase testers, such as elite inbreds, related inbred lines, or single crosses of related lines, making it unnecessary
to conduct preliminary screening tests with heterogeneous testers. This change has evolved because in mature hybrid breeding programs, inbred line development is
directed more toward replacement of a specific inbred line in an already established hybrid, rather than toward the development of a group of new inbred lines to be
used in the production of completely new hybrids. If an inbred line is sought to replace an inbred in an established singlecross hybrid, the opposite inbred would be
the logical tester to use. The testing procedures should be designed so that they will identify whether the new inbred line corrects the weakness of the inbred line being
replaced. Growing the testcross progenies at multiple locations is essential to evaluate genstype × environment interactions and to identify inbred lines with stable
progeny performance in a broad array of environments.
Cytoplasmic Male Sterility and Hybrid Seed Production
A new era in breeding hybrid cultivars was introduced with the identification of cytoplasmic male sterility (cms) and fertilityrestoring genes (Chapter 7), and the
development of procedures for their utilization in hybrid seed production. It was no longer necessary to detassel the seed parent rows when producing hybrid seed
corn, and commercial hybrid seed production became feasible in an assortment of other field and vegetable crops where, previously, restrictions on pollination control
made hybrid cultivars impractical.
Utilization of cms in Hybrid Seed Production
Cytoplasmic male sterility as an aid to hybrid seed production was first utilized commercially in the onion in the late 1940s (Figs. 7.5 and 7.6). A male sterile onion
strain was planted adjacent to a strain with normal pollen production. The male sterile onion strain was pollinated by windblown pollen from an onion strain with fertile
pollen. The F1 hybrid seeds harvested from the male sterile plants were utilized for the commercial production of hybrid onions. Corn and sorghum were the first major
field crops to utilize the cytoplasmicmale
Page 210
sterility:fertilityrestorergenesystem for commercial production of hybrid seed. In corn, the system offset scarcity of labor required for detasseling and reduced costs
associated with the production of hybrid seed. In sorghum, in which male and female floral organs are present in the same flower, the system made hybrid seed
production possible.
The Aline, Bline, Rline Model for Hybrid Seed Production
The onion model for producing singlecross hybrid seed was developed utilizing cytoplasmic malesterile lines called Alines; malefertile maintainer lines called B
lines; and malefertile, fertilityrestoring lines called Rlines. The procedure is accomplished in these steps:
• introduction of a malesterile cytoplasm into the Aline by the backcross procedure,
• maintenance of the malesterile Aline by pollination from a malefertile maintainer line with identical genotype, the Bline,
• development of fertilityrestorer lines called, Rlines, and
• crossing Alines × Rlines to produce hybrid seed.
The Aline, Bline, Rline procedure for onion is illustrated in Figure 11.5.
Normally two or more nuclear, fertilityrestorer genes and some modifying genes are required to restore complete fertility to the cytoplasmic male sterile Aline. The
cytoplasm and a twogene (Rf1, Rf2) fertilityrestorer system are shown for the different breeding lines in the example below:
ALINES AND BLINES. In the cross to produce hybrid seed, the Aline is the female or seedproducing parent line. It contains recessive, nonrestorer (rf1, rf2)
genes. Inbred lines that do not contain dominant, fertilityrestorer genes are crossed to a source of cytoplasmic male sterility and successively backcrossed until the
genotype of the Aline is recovered in the male sterile cytoplasm. The line with the cytoplasmic male sterility is used as the female parent in the crosses because the
cytoplasm is transmitted through the egg, not the pollen. Normally five to seven backcrosses are required to transfer the Aline chromosomes into sterile cytoplasm
and fully recover the genotype of the recurrent parent. The genotype of the Aline must combine with that of an Rline to produce a productive hybrid.
The cytoplasmic male sterile Aline is maintained by pollination from a malefertile
Page 211
Fig. 11.5.
The Aline, Bline, and Rline model for producing singlecross
hybrid seed in onion. The malesterile Aline is maintained by
pollination from a male fertile, genetically identical, nonrestorer
Bline. Hybrid seed is produced by pollination of the Aline from
the fertilityrestoring Rline.
counterpart, called the Bline. The Bline has the same genotype as the Aline, with normal instead of sterile cytoplasm, and recessive, nonrestorer (rf1, rf2) genes
like the Aline.
RLINES. The Rline is the pollen parent in the cross to produce hybrid seed. The function of the Rline is to:
Page 212
• pollinate the Aline,
• restore fertility in the hybrid seed, and
• combine with the Aline to produce a vigorous and productive hybrid.
Abundant pollen production and complete anther extrusion to aid in pollen dissemination are essential characteristics of the Rlines. Because the Rline has dominant
genes for fertility restoration, it may have either fertile cytoplasm or sterile cytoplasm. It may be advantageous to develop restorer lines with sterile cytoplasm, because
presence of restorer genes can then be confirmed without testcrossing to male sterile lines. Two restorer genes are shown in this example, but a third restorer gene and
additional modifying genes are often required to completely restore fertility to the hybrid progeny in different species.
Another requisite of the Rline is ability to combine with the Aline to produce a superior yielding hybrid. The combining ability of potential Rlines is evaluated by
crossing with Aline testers and growing the hybrid progenies in yield trials. If a line with superior combining ability is identified, restorer genes may be added by the
backcross procedure. Modifier genes usually are not added with the backcross but must be present in the potential Rline for it to be a successful parent line.
CROSSING A AND RLINES. The hybrid seed is produced on the Aline after pollination from the Rline. In commercial seed production, the A and Rlines
are planted in alternate strips in the hybrid seed production field, with the male sterile Aline being pollinated by windblown pollen from the Rline. In corn,
sunflower, and wheat, the symbol Rf is used to designate the fertilityrestoring gene. The symbol Ms is used for the fertilityrestoring gene in onion, millet, and sorghum.
An important feature of the cytoplasmicmale sterility/fertilityrestorer gene system is that it made possible the production of hybrid cultivars in crop
species in which male and female floral organs are contained in the same floral structures. Hybrid seed could now be produced in crop plants such as faba
bean, millet, rice, sorghum, sugarbeet, sunflower, wheat, and other field and vegetable crops that, unlike corn, do not have the luxury of detasseling for pollen control.
Alternative Hybrid Procedures
Hand Emasculation and Pollination
Hand emasculation and hand pollination are normally used to make crosses and obtain hybrid seed for growing an F1 generation in the breeding nursery. In most field
crops, hand emasculations and pollinations are too laborious and expensive to use for commercial production of hybrid seed. But there are a few exceptions in crops
where a large number of seeds are obtained from pollinating a single flower, as in cotton and tobacco; or horticultural crops such as tomato, cucumber, melons, and
pepper; or flowers where the hybrid seeds are sold at premium prices.
SelfIncompatibility
Various procedures have been proposed for utilization of selfincompatibility in the production of hybrid seed. The simplest procedure is that utilized with cross
pollinated species
Page 213
that have the gametophytic type of selfincompatibility and that can be clonally propagated. Alternate strips of selfincompatible, yet crosscompatible, clones are
transplanted into an isolated field. Crosspollination among the clones will produce hybrid seeds; selfed and sibbed seed production will be largely prevented by the
selfincompatibility of the clones. All plants grown in the isolated crossing field may be harvested for seed. This procedure was used in the production of the 'Tifhi No.
1' cultivar of Pensacola bahiagrass.
Hybrid seed production in plant species having a sporophytic selfincompatibility system is dependent upon the production of inbred lines homozygous for an S, self
incompatibility allele. In these species, a protein secretion covers the stigmatic surface just prior to anthesis and acts as a barrier to penetration of the stigma by
germinating pollen grains. If buds are opened and the pollen applied before the protein barrier is formed, a procedure called bud pollination, seed set will be
obtained. Bud pollinations are utilized in maintaining selfincompatible parent lines by selfpollination. None of the major field crop plants have the sporophytic self
incompatibility system, but utilization of the system to produce single, double, and threeway crosses in species of Brassica was described in Chapter 6.
Clonal Propagation of an F1 Hybrid
In clonally propagated species, vigorous F1 hybrid plants may be clonally propagated, faithfully reproducing the genotype of the hybrid parent plant. This procedure is
used in breeding sugarcane, potato, forage species such as bermudagrass, and various ornamental, fruit, and forest tree species. Because clonal propagation is
expensive, its utilization in field crops is generally limited to perennial or high value crops such as sugarcane or potato.
Hybrids in Dioecious Species
In perennial, dioecious species, hybrids may be produced by crossing pistillate × staminate clones. When a superior hybrid combination has been identified,
commercial F1 hybrid seed may be produced by interplanting the male and female parent clones in an isolated field. The 'Mesa' cultivar of buffalograss was produced
in this manner.
Hybrids in Monoecious Species
In castor bean, a monoecious species, a plant was found that produced only pistillate flowers. When crosspollinated, it serves as the seed parent for production of
hybrid seed. The pistillate character is controlled by homozygous recessive genes (ff), and when pollinated by a heterozygous monoecious plant (Ff), the progeny
segregates 50% pistillate:50% bisexual plants. Hybrid seed is produced in isolation on pistillate plants by roguing out the bisexual plants.
Apomictically Propagated F1 Hybrids
Apomixis is the production of seed without fertilization (union of gametes). Obligate apomixis is reproduction solely by apomixis, in contrast to facultative
apomixis, in which plants reproduce both sexually, by gametes, and by apomixis. If a vigorous F1 hybrid plant is produced in an apomictic species, and then
converted to an obligate apomict, the genotype of that particular hybrid will be uniformly propagated in all of the progeny. This procedure was used in the production
of 'Higgins' buffelgrass.
Page 214
Genetic Male Sterility
Genetic male sterility is controlled by homozygous recessive genes (msms) (Chapter 7). The effects vary from reduced anther size and pollen production to
complete pollen abortion, according to the species and the specific male sterile gene. Pollen production is restored with the dominant gene, either in the homozygous
(MsMs) or heterozygous (Msms) condition. Producing a completely male sterile population is not feasible, the nearest approach being 50% male sterile and 50% male
fertile, obtained by pollinating a malesterile plant with pollen from a plant heterozygous for the malesterile gene. This restricts the utilization of genetic male sterility in
the commercial production of hybrid seed in crop plants where large quantities of seed are required, because the 50% malefertile plants would need to be rogued out
before they shed pollen, a laborious and expensive hand procedure. Systems for hybrid seed production using genetic male sterility have been proposed for use in
barley, corn, sorghum, and wheat, but none has proved practical.
Chemically Induced Male Sterility
Chemical induction of male sterility would eliminate the need for developing cytoplasmic malesterile and fertilityrestoring lines in the commercial production of F1
hybrid seed. The chemical hybridizing agent is sprayed on the seed production rows in hybrid seed production during a critical stage of flower formation to induce
male sterility. The potential exists for efficiently sterilizing large numbers of breeding lines to obtain F1 seed for yield evaluation trials. Any pair of lines would become
potential parents in a hybrid if they combined to produce a highyielding F1 hybrid, and one of the lines could be successfully sterilized and used as the seed parent.
Consistency in inducing male sterility in widely varying environments is essential. Because the chemical must be applied at a critical stage of floral development, some
risk is involved if application of the chemical is interrupted by adverse weather conditions. Chemical sterilization for hybrid seed production has been proposed for use
in cotton and wheat.
Proprietary Nature of Hybrid Cultivars
An important consideration in the development of hybrid corn and the extension of hybrid breeding to other crops has been the proprietary nature of the hybrid
cultivar. When hybrid corn was first introduced to the U.S. farmer, the inbreds utilized in the early hybrids had been developed by scientists in the state agricultural
experiment stations and the U.S. Department of Agriculture. As taxsupported research organizations, it was not appropriate for either one to produce and market
hybrid seeds. That activity was left to the private hybrid seed companies that were rapidly emerging. As the resources of the private seed companies increased from
sales of hybrid seeds, they developed breeding programs. At that time, legal protection of the intellectual property rights of developers of new cultivars did not exist in
the United States, but a hybrid cultivar has a builtin protection for the developer if the developer maintained sole possession of one or more of the inbred lines. This
proprietary protection encouraged investment by private seed companies in hybrid corn research and in inbred line development. In the United States hybrid corn is
produced and distributed by private seed companies. Hybrid breeding was later extended to sorghum, pearl millet, sunflower, wheat, and other field and horticultural
crops.
Page 215
How Breeding Procedures Are Utilized
In this chapter we have described how hybrid cultivars originated and discussed problems associated with their development and utilization. For a discussion of
specific procedures and problems in breeding hybrids, the student is referred to Chapter 17, Breeding Corn, and Chapter 18, Breeding Sorghum.
Study Questions
1. What are the consequences of inbreeding? What are the various methods that the plant breeder can use to obtain inbred plants?
2. What theories are used to explain hybrid vigor? What are the pros and cons for each theory?
3. What is general combining ability and specific combining ability? Why is it necessary to know which type of combining ability is most important for a particular crop
species?
4. Why are double crosses no longer used in the breeding of corn?
Further Reading
Hayes, H.K. 1952. Development of the heterosis concept. p. 4965. In J.W. Gowen (ed.) Heterosis. Iowa State Univ. Press, Ames, IA.
Hayes, H.K. 1963. A professor's story of hybrid corn. Burgess Publ. Co., Minneapolis, MN.
Jones, D.F. 1918. The effects of inbreeding and crossbreeding upon development. Conn. Agric. Exp. Stn. Bull. 207.
Jones, H.A., and A.E. Clarke, 1943. Inheritance of male sterility in the onion and the production of hybrid seed. Proc. Am. Soc. Hort. Sci. 43:18994.
McRae, D.H. 1985. Advances in chemical hybridization. p. 16991. In J. Janick (ed.) Plant Breeding Reviews, Vol. 3. AVI Publ. Co., Westport, CT.
Shull, G.H. 1908. The composition of a field of maize. Rep. Am. Breeders Assoc. 4:296301.
Shull, G.H. 1909. A pureline method in corn breeding. Rep. Am. Breeders Assoc. 5:5159.
Shull, G.H. 1952. Beginnings of the heterosis concept. p. 1448. In J.W. Gowan (ed.) Heterosis. Iowa State Univ. Press, Ames, IA.
Sprague, G.F., and J.W. Dudley (eds.). 1988. Corn and corn improvement. 3rd ed., Agronomy Monograph 18. Am. Soc. Agron., Crop Sci. Soc. Am., and Soil Sci.
Soc. Am., Madison, WI.
Sprague, G.F., and L.A. Tatum. 1942. General vs. specific combining ability in single crosses in corn. J. Am. Soc. Agron. 34:92332.
Sprague, G.F., and S.A. Eberhart. 1977. Corn breeding. p. 30563. In G.W. Sprague (ed.) Corn and corn improvement, 2nd ed., Agronomy Monograph 18. Am.
Soc. Agron., Crop Sci. Soc. Am., and Soil Sci. Soc. Am., Madison, WI.
Stephens, J.C., and R.F. Holland. 1954. Cytoplasmic malesterility for hybrid sorghum seed production. Agron. J. 46:2023.
Page 216
12. Breeding Objectives and Techniques
The goal of the plant breeder, whether developing a cultivar or a parent for a hybrid, is to create new genotypes improved in one or more important features. The goal
is achieved only if the breeding objectives are clearly defined, and carefully planned selection and hybridization procedures leading toward that goal are employed. The
breeder must identify the changes in the cultivar which, if made, will increase yield, stabilize production, or improve quality of the product to be harvested
in the crop species with which he is working. He then searches for parent materials superior in those features and combines genes for the desired features
into new cultivars to fulfil the breeding objectives. The objectives in breeding programs will differ with the crop species and the purpose for which the crop is
being grown. The objectives for a particular species will differ from program to program because the environmental conditions that affect production and the
adversities that limit yield differ from one production area to another. By designing cultivars to fit his particular concept of the ideal cultivar, the experienced breeder
leaves a lasting imprint on the cultivars that he develops.
This chapter presents an overview of the major objectives in crop breeding and the breeding techniques by which plant genotypes are screened to identify whether the
objective has been accomplished. The relative importance of the objective may differ with the crop species as will be discussed in chapters concerned with the
breeding of different species. A prudent choice of objectives is essential for a successful breeding program; since the breeder cannot pursue all possible objectives he
must concentrate available resources on those objectives most important in his particular breeding program.
Yield
Yield of grain, forage, fiber, or other plant products has primary importance as a breeding objective because it affects the economic return to the grower. Plant
genotypes differ in their inherent yield potential. Yield potential is expressed phenotypically through complex plant morphological features and physiological functions,
and genetically expressed as a complex
Page 217
quantitative character that interacts with the environment in which the plant genotype is grown. The plant breeder measures yield potential by the mass or weight of the
product produced per unit of land area, through yield trials, in which the harvested yield is compared with that of standard cultivars.
Breeding for highyield potential is normally accomplished by crossing among genotypes with complementary genes contributing to yield potential in order to
generate transgressive segregates with superior yield. In the early, segregating generations single plants are selected, which are then evaluated for yield potential by
progeny tests. To identify the superior segregates, breeders are faced with the dilemma that for the initial selection, they must rely on visual observation among
thousands of spaced plants or progeny rows, yet the critical test is how the genotype performs at normal spacings in yield trials and in farmer's fields, in comparison
with the highest yielding commercial cultivars.
Various attempts have been made to find selection criteria that will aid the breeder in selecting for yield during the early selection stage, but none are entirely
satisfactory. One approach has been to examine the components of yield. For example, yield of wheat is the product of number of spikes per unit of land area,
number of grains per spike, and average weight per grain. In theory, an increase in one component with the others held constant would result in an increase in total
yield. In practice, as one component of yield is increased, the others tend to decline due to competition for available growth assimilates. Among .the different yield
components, kernel weight tends to have the highest heritability coefficient.
With a combination of shorter stature and higher yield in new cereal cultivars, there has been an increase in the ratio of grain weight to total plant weight (grain +
straw), the ratio being designated harvest index. A higher harvest index is an expression of greater plant efficiency since a higher proportion of the available assimilates
is being translocated into the grain rather than being used to produce plant mass. This has led to suggestions for increasing yield by selection for harvest index. The
problem here is that harvest index is not a phenotypic plant character that can be visibly selected. Harvest index is a ratio, its calculation requires separate weights of
grain and straw, information that is impractical to obtain with modern mechanized harvest of yield trials.
How, then, does the breeder select for highyield potential in the early segregating generations? Most selection for yield potential at this stage is empirical, based on the
breeder's knowledge and experience, and the accuracy of his observations. Uniform guidelines for selection are difficult to establish. They differ with the crop species
and its intended use, the breeder's experience, and the local environment. Because the breeder must examine large numbers of plants in the early generations, selection
criteria that require detailed measurements are generally impractical. Selection at this stage must be followed by extensive field testing at various locations in the area
where the new cultivar will be grown.
All this assumes that the cultivar has a favorable environment in which to grow; that adverse factors such as cold, drought, or disease will not limit the final yield. To
find such an environment would be rare indeed. And so in addition to highyield potential, we breed for yield stability, the ability of the plant genotype to produce up
to its genetic potential in spite of an adverse environment. We breed for cold resistance in areas where nonhardy plants would be injured, for resistance to drought in
the dry areas, for sturdy straw to prevent loss from lodging, for resistance to soil stress in the presence of excess aluminum or toxic salts, or for resistance to disease
pathogens and insect pests that affect the plant's health. While constantly striving to improve the potential yielding ability, it is also necessary to stabilize production by
breeding for resistance to the adversities that may limit the final harvest.
Substantial increases in grain yields for wheat, rice, sorghum, and hybrid corn have been
Page 218
recorded over the past 30 to 50 years, particularly in highyield environments as in the United States, Western Europe, India, or China. The higher yields result from
genetic improvement in yield potential; shorter and stronger plants; and resistance to stress, disease pathogens, and insect pests; combined with improved cultural
practices such as higher rates of fertilization; increased use of chemicals for control of weeds, disease, and insect pests; and heavier seeding rates to give more plants
per unit of land area. The interaction of the new cultivars with the improved cultural practices is complementary. Because the new cultivars have shorter and stronger
stems, increased fertilizer applications may be made without lodging; and with larger inherent yield response to increased soil amendments, greater use can be made of
irrigation, herbicides, fungicides, and insecticides to prevent yield loss from adverse circumstances.
Maturity
The optimum period for a cultivar to reach maturity will be altered by where the cultivar is grown; its place in the cropping sequence; the importance of maturity in
escaping disease, insects, or other natural hazards; and the relation of maturity to yield and product quality. There has been a continuing trend toward earlier maturity in
cultivar development. An early maturing cultivar may escape damage from heat, drought, insects, or disease, or permit harvest ahead of damage from rainstorms, hail,
floods, or early frosts. Early maturity is advantageous in multiple cropping systems to permit early removal of a crop so that the following crop may be planted. There
are also disadvantages in early maturity. Plant size and yield may be reduced in extremely early cultivars of crops such as corn, soybean, or wheat, because the plant
has a shorter growth period in which to develop, manufacture, and store nutrient materials. Early maturity may be the result of early flowering or a reduced time period
from flowering to ripening. Environmental factors such as photoperiod, temperature, altitude, soil type, and seasonal distribution of moisture affect maturity as well as
the plant genotype.
The comparative maturity of crop cultivars are expressed in various ways, some of the more common being daystoheading (small grains), daystosilking (corn), or
daystoripening. In small grains, daystoheading is influenced less than daystoripening by abnormal temperatures, drought, or disease, environmental factors that
cause premature ripening. Because all plants within a plot do not flower on the same date, estimates of the date that 75% of the plants flower may be used in
comparisons of maturity. The moisture percentage of the grain at time of harvest provides a measure of relative maturity in corn. In cotton, earliness may be measured
by daystofirstflower, length of the bollforming period, or percentage of lint at first harvest. A standard cultivar is included in all field trials for comparison.
The inheritance of earliness differs with specific crops and cultivars, but is frequently reported as being dominant or partially dominant to late maturity and controlled by
a few major and numerous modifier genes. In wheat chromosome substitution studies, seven or more chromosomes are reported to have genes affecting maturity.
Maturity is affected by photoperiod response which is controlled by relatively few genes.
Resistance to Lodging and Shattering
Lodging is the bending or breaking over of the plants before harvest. Lodging causes yield losses in small grains, soybean, corn, sorghum, and other crops, with any of
the following conditions:
Page 219
• the plant lodges before it is ripe and the seeds do not fill properly,
• the fallen plant is not picked up in the harvest operation, and
• the fallen plants provide a favorable environment for the development of disease (rust, mildew) or for multiplication of insect pests.
Rain, hail, and windstorms occurring after plants have flowered, but before they ripen, are common causes of lodging in small grains. The plants at this stage are green
and heavy and are easily broken over by the added weight of the rain or the force of the wind (Fig. 12.1A). Disease and insect injury that weaken the stem or other
causes of lodging in small grains (Fig. 12.1B and C) and other crops.
Breeding to improve resistance to lodging involves changing the architecture of the plant, including the root system, to improve its structural strength and stability, and
acquiring resistance to diseases and insects that weaken the plant structure. Resistance to lodging may be improved by the development of cultivars with:
• short, stiff, sturdy stems or stalks,
• vigorous, strong root systems that anchor the plant firmly in the soil,
• more resilient straw that will bend but not break in the wind, and
• resistance to disease and insects that weaken the stem or the root system.
Tall plants with slender or weak stems, as in soybean, or plants that are succulent as a result of excessive nitrogen fertilization or soil moisture, are most susceptible to
lodging. In crops such as wheat and rice, dwarfing genes have been utilized to reduce height and increase lodging resistance.
Fig. 12.1.
Comparison of lodging in wheat cultivars. (A) Lodging
from wind and rainstorm before the wheat had ripened.
(B) Lodging due to disease. Straw of the susceptible cultivar
was weakened by stem rust. (C) Lodging from insect injury.
The broken and fallen culms of the susceptible cultivar
were injured by infestation from the Hessian fly.
Unlike the measurement of yield, in which the quantity is physically measured, evaluation of lodging resistance is almost entirely a visual appraisal. It is obtained by
comparing the
Page 220
relative amounts of bending or breaking over of cultivars growing in adjacent nursery or field plots. It is necessary that all cultivars be grown under as nearly identical
conditions as possible, and that a standard cultivar be uniformly included to which experimental strains or new cultivars may be compared. Lodging of an intensity to
permit accurate differentiation of cultivars or breeding lines does not occur consistently, making it necessary to replicate field trials at different locations and over a
period of seasons. Lodging may be intensified by liberal applications of nitrogen fertilizers, or by letting portions of the yield nursery stand for a period after ripening.
Visual observations on lodging in small grains are normally recorded on a percentage basis (zero to 100), or on a scale of 1 to 10. In com hybrids, lodging may be
expressed as percentage of plants with root lodging (leaning more than 30% from the vertical), or with stalk breakage (stalk broken below the ear) (Fig. 12.2). The
structural strength of a corn stalk may be measured by the force required to crush a section of the stalk in a hydraulic press. About 50 to 70% of the strength,
designated crushing strength, is in the outer rind and the remainder is in the pith. Lodging resistance in corn was increased by recurrent selection for increased
crushing strength. The firmness with which plants of different corn hybrids are anchored may be measured by the force required to pull the plant, designated the root
pulling force. Lodging resistance is a quantitative character with complex inheritance, although some of the plant characteristics associated with a reduction in lodging,
such as dwarfing genes, or resistance to disease and insect pests, are often simply inherited.
SHATTER RESISTANCE. Shattering refers to seeds that fall out and are lost before harvest or during the harvesting operation. Resistance to shattering is
important to prevent loss of yield in small grains, soybean, and some other crops. Visual estimates of loss are commonly made to compare the resistance of cultivars of
wheat, soybean, or other crops to shattering. Resistance to shattering is inherited as a complex quantitative character.
Fig. 12.2.
Types of lodging in corn, Lodging in corn is generally expressed as root lodging,
if the stalk leans more than 30 degrees from the vertical, or stalk breakage,
if the stalk is broken below the ear.
Page 221
Winter Hardiness
Winter hardiness refers to the capability of crop plants to survive severe winter stress. The most common types of winter injury to crop plants are:
• freezing of plant tissues from the combined effects of low temperature, wind, and insufficient soil moisture, and
• heaving or uplifting of plants, shearing off the roots as the plant is torn from the soil, caused by alternate freezing and thawing in water saturated soils.
Plant genotypes differ in resistance to freeze and heaving injury. Winter hardiness is a complex character, the expression of which is affected by environmental and
cultural factors such as (a) low temperature, (b) wind, (c) hardening of the plant, (d) alternate freezing and thawing, (e) moisture content of the soil, (f) physical
condition and fertility of the soil, (g) time and rate of planting, (h) disease and insect injury of the plant, and (i) snow cover. Field crops most generally subjected to
winter injury include the winter cereals (wheat, barley, oat, rye), alfalfa, rape, and various root and forage crops.
Low temperature survival depends upon the plant's inherent ability to harden when exposed to temperatures around 5ºC, and is affected by the rate of freezing and
thawing, light, and nutrient status and health of the plant. Low temperature injury is associated with ice formation in the extracellular spaces resulting in freezeinduced
dehydration and complex metabolic changes. With low soil moisture and a high windchillindex, the plant tissues become desiccated. Cold resistant plants are
generally drought resistant, also. Resistance to low temperature injury varies with the species and with genotypes within the species. In the small grains, rye is the most
cold hardy species followed by, wheat, barley, and oat in that order. Within a species, the most cold hardy cultivars have generally originated from germplasm that
evolved in a severe climatic region. Examples are 'Turkey Hard Red Winter' wheat from the Crimea; and 'Grimm' alfalfa that was subjected to intense natural selection
in Minnesota after its introduction from Germany. Cultivars and breeding lines are evaluated for winter hardiness by growing in field trials in areas where they will be
subjected to winter injury (Fig. 12.3). Standard cultivars with varying levels of winter hardiness are included for comparison. Survival scores are recorded on a scale
of 1 to 100, or a scale of I to 10, from visual observation. Laboratory procedures for measuring cold hardiness are usually of limited utility because they measure only
one form of winter stress. Because disease or insect damaged plants are injured more severely than healthy plants, loss from winter injury may be reduced by breeding
for resistance to disease pathogens or insect pests causing damage to fallseeded crops.
Heaving occurs on heavy, moisture saturated soils. Ice crystals accumulate following alternate freezing and thawing, forming successive layers of ice lenses, and
uplifting the plant. The heaved plant is frequently uprooted and left on the soil surface where it dies from desiccation. Inherent resistance to heaving is associated with
the development of a dense and healthy root system that anchors the plant firmly in the soil. Heaving is more severe in crops with tap roots than with a branching root
system.
Page 222
Fig. 12.3.
Comparative winter survival of four strains of winter oat at Columbia, Missouri. Row A
is the most hardy. Row B is less hardy than A. Row C is completely killed. Row
D has one surviving plant.
Heat and Drought Resistance
Damage and yield loss caused by heat and drought stress are common in hightemperature, lowrainfall areas. The damage may be increased if accompanied by the
desiccating effect of strong winds. The effects of heat stress will vary with the crop species, the stage of plant growth, and the duration of the stress and are
confounded with the effects of drought stress. Because respiration increases at a faster rate than photosynthesis with increases in temperature, photosynthate supply
becomes exhausted, suppressing growth and reducing yield. Injury from heat stress is critical during the flowering period reducing pollen viability, stigma receptivity,
and seedformation. Breeding for earlier maturity may sometimes permit flowering to occur before periods of high temperatures.
Drought resistance mechanisms generally relate to drought avoidance or drought tolerance. Drought avoidance mechanisms include deeproot systems to acquire
soil moisture at lower levels, and plant characteristics to reduce water loss, such as closed stomata, leaf rolling, or waxy substances on the leaf surface. Usually heat
and drought tolerance are considered together because high temperatures and drought often accompany each other in the field. Soil moisture gradients from irrigation
lines have been used to evaluate strains in different levels of drought stress. Irrigation water supplied will be highest near the irrigation line and will decline outward until
a point is reached where no irrigation water is received. Cultivars growing along the gradient are compared for performance.
There is a high correlation between tolerance to heat and tolerance to desiccation. Comparative resistance of plant genotypes may be observed by exposure to high
temperature, soil drought, or atmospheric drought. Like winter hardiness, plant genotype resistance to heat and drought stress in the field is determined by complex
physiological and morphological characteristics and cannot be accurately evaluated by a single laboratory test. Heat and drought stress resistance are quantitative
characters with complex inheritance.
Page 223
Soil Stress
Soil stress refers to problem soils caused by aluminum or saline toxicity, or various acid or alkaline soil problems. Because many soil problems are related to pH, or
other interrelated factors, the precise problem needs to be identified and screening tests devised to evaluate genotype resistance to the specific problem. Screening
tests for aluminum or saline tolerance will be more precise if conducted under controlled experiments than when conducted in the field because concentration of the
toxic element may be controlled more accurately.
Aluminum Tolerance
Barley and wheat cultivars developed in regions of the world with acid soils and high soluble aluminum usually have higher levels of aluminum tolerance than cultivars
developed in regions with nonacid soils. High aluminum soils tend to restrict root and shoot development. Cultivars and breeding lines may be screened for aluminum
tolerance in the laboratory by growing seedlings in a nutrient solution containing a high concentration of aluminum ions and selecting for plants with longest root and top
growth (Fig. 12.4). When grown in the field on aluminumtoxic soils, aluminumtolerant cultivars survive winter or drought stress and produce higher yields than
aluminumstressed cultivars, due to the larger root development and top growth.
Salt Tolerance
Crop species are nominally sensitive to high salt concentrations such as is present in sea water. Because large soil areas contain high salt concentrations, efforts have
been made to generate crop genotypes with tolerance to this condition. Barley has one of the higher levels of tolerance to salt concentration among the major crop
species and studies conducted with this species demonstrated that salt tolerance is a heritable trait. In tomato, salt tolerance was identified in a related wild species and
transferred to cultivated tomato.
Resistance to Plant Disease Pathogens
Cultivars of crop plants developed with genetic resistance to destructive disease pathogens are among the foremost contributions in crop breeding. In breeding for host
resistance to
Fig. 12.4.
Effect of aluminum toxicity on root systems of nontolerant
and tolerant strains of barley.
Page 224
different crop diseases each disease is treated as a separate breeding objective. Because crop species are attacked by many disease pathogens, the breeder must
establish priorities and concentrate available resources on developing cultivars that are resistant to the most destructive diseases prevalent in the area where the cultivar
will be grown. Disease resistant cultivars are developed by identifying genes for resistance in the host species, or related wild species, and transferring the gene or
genes into adapted cultivars and breeding lines, normally by hybridization or chromosome engineering techniques. As molecular biology develops, it should become
possible to reach out and move genes for resistance from sources not possible through hybridization techniques. Resistance may be controlled by single genes or by
polygenes depending on the specialization of the pathogen and the nature of the resistance.
Breeding for host resistance to particular diseases is complicated by the presence of physiologic race specialization in the disease pathogen. If physiologic race
specialization is present, various races or biotypes of the pathogen are present in the pathogen population. Resistance in the host genotype to specific races, or blocks
of races, of the disease pathogen is conferred by racespecific genes. Racespecific genes are simply inherited and confer major resistance effects to particular races
or biotypes of the pathogen but not to other races. In highly specialized disease pathogens, many physiologic races may be present, each controlled by a different
racespecific gene.
In the absence of race specialization in the pathogen, resistance in the host genotype is conferred by nonracespecific polygenes. Nonracespecific polygenes are
inherited quantitatively, each contributing a small increment of control of the disease pathogen. Reduction in disease damage may also be achieved by breeding for
plant characteristics that enable the plant to escape or avoid disease infection.
Breeding for RaceSpecific Resistance
A familiar example of a highly specialized plant pathogen is the pathogen inciting the stem rust disease (Puccinia graminis f. sp. tritici in wheat and P. graminis f. sp.
avenae in oat). P. graminis is composed of numerous specialized biotypes, or physiological races; each race constitutes a population of rust fungi that is genetically
different from populations of other races, just like a wheat cultivar is genetically different from other wheat cultivars. The rust fungal race may possess a gene that is
virulent (capable of infecting) on a specific wheat genotype, yet be avirulent (incapable of infecting) on another wheat genotype. The host genotype (cultivar or
breeding line) may possess a gene, or genes, for resistance to one race of the pathogen, yet be susceptible to another race of the pathogen (Fig. 12.5). The interaction
between the gene for virulence in the pathogen and the gene for resistance in the host plant determines whether the reaction is designated resistance or susceptibility.
New physiologic races arise through hybridization among biotypes of the pathogen, or by mutation. If a new cultivar with a gene for resistance to a widespread race of
the pathogen is distributed, the new cultivar will screen the avirulent races out of the pathogen population. If a new race of the pathogen arises that is virulent on the
new cultivar, the new race may multiply and becomes the prevalent biotype. The breeder must then identify a gene for resistance to the new race and introduce it into
the cultivar or breeding lines.
The pathogens inciting the cereal rust diseases are especially notable for their physiologic specialization, but physiological race specialization is found in a host of other
disease pathogens and insect pests, including the pathogen inciting powdery mildew, and the insects, Hessian fly and greenbug. Racespecific resistance is normally
conferred by a single dominant gene, but may be conferred by two or more genes, or by recessive genes, normally with large effects.
Page 225
Fig. 12.5.
Reaction of an oat cultivar, 'MO. 0205', to two races
of Puccinia graminis avenae. Left: Stem showing resistant
reaction to physiologic race no. 7. The resistanttype pustules
appear as small black dots. Center: Stem showing susceptible
reaction to physiologic race no. 8. The large conspicuous
pustules have ruptured the epidermal layer. Right: Stem infected
with physiologic races 7 and 8. Both small, resistanttype pustules and
large, susceptibletype pustules are present.
Cultivar life is generally thought to be extended by pyramiding genes—the practice of introducing several genes for resistance into a cultivar, each of which would need
to be overcome separately by a new race of the pathogen. Reaction to racespecific pathogens may range from immunity, with no visible sign of infection, to
hypersensitivity, in which the host cells immediately around the infection site are killed preventing further spread of the disease pathogen, to susceptibility, in which a
large sporeproducing pustule develops (Fig. 12.6).
Breeding for NonRaceSpecific Resistance
The rapid development and spread of new physiologic races in some disease pathogens and insect pests has necessitated a rapid turnover in cultivars to keep pace
with the changes in the parasite. In breeding for resistance to the crown rust pathogen, Puccinia coronata, in oat, new cultivars with different genes for resistance
were introduced in 1940, 1946, and again in 1953 to replace previously resistant oat cultivars that had succumbed to new races of the crown rust pathogen.
Accumulated experiences with specialized pathogens such as the rust fungi have demonstrated that:
• high levels of host resistance to a specific race or a combination of races of the pathogen may be conferred by a single, dominant, racespecific gene, that restricts
the infection process as described above,
• the useful life of cultivars with resistance conferred by racespecific genes may be short as new virulent races of the pathogen arise and become widespread, and
• resistance conferred by nonrace
Page 226
Fig. 12.6.
Types of reaction to the stem rust pathogen, Puccinia graminis tritici,
on wheat cultivars. (A) Immune, no visible infection. (B) Hypersensitive,
the host cells are killed immediately around the infection site.
(C) Susceptible, large spore filled pustules.
specific polygenes that resist colonization and spread of the pathogen would be less vulnerable to damage by new races of the pathogen.
Various nonracespecific systems have been described and designated as complete resistance, durable resistance, general resistance, partial resistance, or slow
rusting. Nonracespecific genes for resistance do not interact differentially with pathogen races, the polygenic resistance extending to all physiologic races. The
effects of the polygenes for resistance are small, generally expressed by slowing the rate of disease development through reduced spore production or by other means.
The inheritance is complex, improvement being achieved through transgressive segregation.
Avoidance
Avoidance refers to breeding for plant characteristics to escape disease or insect damage. An example is the use of early maturity to ripen wheat cultivars ahead of
stem rust epidemics in the U.S. Great Plains.
Techniques for Inducing Disease Epiphytotics
In breeding for disease resistance, exposure to the disease pathogen, either in natural or artificially induced epiphytotics, is necessary to distinguish between resistant
and susceptible genotypes. Progeny tests of resistant plants are made to verify the inherent nature of the resistance and to ensure that uninfected plants have not merely
escaped infection. A basic problem in breeding for disease resistance is that of providing a disease environment in which to grow the crop so that the resistant plants
may be distinguished from the susceptible. Because natural disease epidemics do not occur in the fields every year, it is desirable for the breeder to introduce disease
epiphytotics by artificial means, either in the field or in the greenhouse, so as not to be entirely dependent upon the vagaries of nature to provide an adequate disease
Page 227
environment. Field trials have the advantage that the host genotype will be tested against pathogen races prevalent in the field. But greenhouse tests have the advantage
of testing against specific races of the pathogen and provides precise control of temperature and humidity for optimum disease development. Close cooperation
between the plant pathologist and the breeder is essential to ensure exposure to the desired biotype of the disease pathogen and selection for both pathogen resistance
and agronomic characters. Cultivars with known reaction, both resistant and susceptible, to the biotypes of the pathogen should be included as checks.
INOCULATION TECHNIQUES FOR SOILBORNE DISEASE. Diseases incited by soilborne pathogens that enter the host plant through roots or other
underground parts, may be conducted by growing the host genotype in soils in which the diseaseinciting pathogen is naturally prevalent, or by growing in controlled
environments in sterilized soil which has been inoculated with cultures of the causal organism (Fig. 12.7). The latter differentiates breeding lines more precisely than
field tests because temperatures favorable for the growth and development of the diseaseinciting organism can be maintained. A cold test for corn has been
developed in which seeds are germinated in contact with soil from fields infested with seedlinginfesting pathogens. The temperature is maintained near the optimum for
the development of the pathogens, which is normally below the optimum for germination of the corn. This simulates early spring planting of corn in cold, wet soils.
Massscreening of seeds for soilborne organisms may be made by spraying germinating seeds with a suspension of the diseaseinciting pathogen. Normal, healthy
seedlings are transplanted to soil and grown to maturity; diseased seedlings are discarded.
INOCULATION TECHNIQUES FOR FOLIAGE DISEASES. Many diseaseinciting organisms infect the foliage by entering through natural openings, such as
stomata or lenticels, or through wounds inflicted during tillage, by insects, or other means. With foliage diseases, such
Fig. 12.7.
Sorghum cultivars tested for resistance to Periconia rootrot or milo disease
by germinating seeds in soil infested with the pathogen, Periconia circirata.
Lines 2, 4, 6, 8, and 10 are resistant. Lines 1, 3, 5, 7 and 9 are susceptible and
have been killed by the disease.
Page 228
as the cereal rusts, dry spores are collected from diseaseinfected plants and dusted on the foliage of plants to be inoculated, or a suspension of spores may be
sprayed on the foliage (Figs. 12.8 and 12.9). The intensity of the disease infection is increased by growing the host plants in a controlled environment with temperature
near optimum for the growth and multiplication of the pathogen and the atmosphere maintained near saturation. Infection may be increased by spraying when the
stomata are open, and by reducing the surface tension of the spore suspension with a mild detergent to obtain an even spread on the leaf surface. A power sprayer or
a hypodermic syringe may be utilized to apply bacterial inoculum in order to force the inoculum into open stomata, or cause water soaking of the leaves to aid entry of
bacteria. Susceptible cultivars planted adjacent to test plants in the field and inoculated with the pathogen serves as a secondary source of infection to the test plants.
Host plants may be tested for reaction to two or more diseases simultaneously by sequential inoculations on successive leaves as new leaves emerge.
INOCULATION TECHNIQUES FOR FLORALINFECTING DISEASES. Inoculations for floral infecting diseases, such as the loose smut of wheat and
barley, are made by injecting ripened spores into the flower during anthesis. Dry spores may be introduced with forceps or a hypodermic needle; or a spore
suspension may be injected into the flower with a hypodermic needle, or by means of a vacuum or pressure. Seeds produced in the inoculated flowers are harvested
and planted, and the percent of infected plants recorded.
INOCULATION TECHNIQUES FOR SEEDBORNE DISEASES. Inoculation techniques for seedborne diseases are made by applying spores from the
pathogen on the seed before seed germination. With common bunt of wheat or covered smut of sorghum, dry spores are dusted on the seeds before planting. If the
seed is covered with a hull, as with oat or barley, the seeds are soaked in a spore suspension under a vacuum. The vacuum withdraws the air from under the hull,
permitting penetration by the spore suspension.
Fig. 12.8.
Collecting spores from a rustinfected plant of wheat by suction.
The rust spores are used to inoculate healthy plants of cultivars
and breeding lines to establish their reaction to races of the stem
rust pathogen, Puccinia graminis tritici.
Page 229
Fig. 12.9.
Dusting rust spores on wheat plants in an incubation chamber.
The plants are sprayed with water to obtain good dispersal
of the spores and kept in a warm humid atmosphere to
foster disease development.
Some diseaseinciting pathogens, such as Gibberella spp. and Diplodia spp., that incite root, stalk, and ear rots of corn, are both seedborne and soilborne.
Inoculation with these pathogens may be made as for soilborne organisms, or by injecting the pathogen directly into the stalk tissue with a hypodermic needle.
INOCULATION TECHNIQUES FOR VIRUS DISEASES. Virus diseases are transmitted mechanically or by employment of insect vectors, depending on the
particular virus. For mechanical transfer of the virus, diseased plant tissue is macerated and the extracted juices are rubbed over the leaves of healthy plants with
sufficient force to cause slight mechanical injury. A fine abrasive such as carborundum powder dusted over the leaves first, or mixed with the juices, will aid in
obtaining injury and providing avenues for uptake of the virus particles. With insect transmitted viruses, insects of the species specific for natural transmission of the
virus feed on virusinfected plants to acquire the virus, and transmit the virus when they feed on a healthy plant. Healthy plants of the genotype to be tested are
inoculated by caging with viruliferous insects that feed on and transmit the virus to the test plants. Plants are grown in insectproof cages to prevent escape of the
insects vectoring the virus, or to prevent feeding on healthy plants by naturally occurring vectors (Fig. 12.10).
Resistance to Insect Pests
Hostplant resistance is utilized to control particular insect pests that may be difficult to control through cultural practices or use of pesticides. Breeding for hostplant
resistance is
Page 230
Fig. 12.10.
Comparative reaction of mungbean strains to a strain
of cucumber mosaic virus (CMV). The CMV is vectored by
naturally occurring cowpea aphids. Mungbean strains grown
in insectproof cages in the rear are virusfree and may be
compared with the virusinfected strains.
environmentally safe as it reduces the employment of harmful pesticides. The principles in breeding for resistance to an insect species and the techniques employed are
similar to those used in breeding for resistance to disease pathogens. Genes for resistance to the insect species are first identified in crop cultivars or related wild
species and transferred to susceptible host genotypes by hybridization. Cultivars and breeding lines are exposed to natural insect populations in the field, or to
artificially reared insect populations in controlled environments, in order to distinguish between resistant and susceptible genotypes of the host species (Fig. 12.11).
Biotypes, or races, have been identified in particular insect species that are comparable to biotypes or races in pathogens. The biotypes are identified by reaction to
insect feeding, whether resistant or susceptible, on cultivars with known genes for resistance. Insect damage is often related to the stage of growth and development of
the plant, so uniform maturity is essential when comparing a series of cultivars or breeding lines for reaction to insect feeding. Cooperation between the entomologist
and the plant breeder is essential in development of an insect resistant cultivar.
Product Quality
The objectives for the improvement of crop plants discussed thus far have been directed toward enhancement of yield potential and yield stability. In the present
market economy, product quality has become increasingly important. In breeding for improved quality, consideration is given to the physical and chemical
characteristics of the product harvested that affect its
• nutritional value,
• processing, and
• utilization.
The characteristics of product quality will vary with the crop species and its intended use. Wheat, a food crop that is processed for many complex uses will be
considered as an example.
Page 231
Fig. 12.11.
Comparative damage to corn (maize) inbred lines
by the corn earworm, Heliothis zea. (above) Susceptible
inbred line. (below) Resistant inbred line.
The principal use of wheat is for milling into flour for making bread, but wheat is also milled into flour for making cakes, cookies, crackers, pretzels, and semolina
products. The different wheat products require genetically different kinds of wheat, grown in different climatic areas, and processed by different milling procedures.
The milling procedure and the end product of each are affected by the genotype of the wheat cultivar, and are subject to modification in the breeding program.
Likewise, other crop species each have distinct quality requirements, that are subject to genetic modification. Developing cultivars with superior quality characteristics,
is an important objective in any crop breeding program.
Techniques in Plant Hybridization
Selfing and crossing are essential procedures in breeding crop plants. The exact procedures employed will depend upon the crop species, the structure of the floral
organs, and the normal manner of pollination.
Selfing
In selfpollinated species, the plant is permitted to follow its normal mode of pollination, although recognizing that some natural crosspollination will occur. If slight, the
cross
Page 232
pollination may be ignored in normal breeding procedures, but if excessive, it may be necessary to protect the flower from foreign pollen by bagging to prevent foreign
pollen from reaching the stigma. In selfing or inbreeding a naturally crosspollinated species, the flower must be covered and protected from foreign pollen. In plants
like cotton, with large flowers, the petals may be folded down over the sexual organs and fastened to exclude pollen and pollencarrying insects. In corn, a paper bag
is placed over the tassel to collect pollen, and the shoot is covered to protect it from foreign pollen. Pollen collected in the tassel bag is transferred to receptive silks on
the same plant to obtain selffertilization. Bagging and handtripping is necessary in legumes such as alfalfa to obtain selfpollination, or the legume plants may be caged
with an insect pollen vector. Many insectpollinated species are highly selfsterile and selfed seeds are difficult to obtain.
Emasculation Practices
Cross pollinations in species with bisexual flowers are generally accomplished by removing the stamens before pollen is shed from the female parent and transferring
pollen collected from the pollen parent to the emasculated florets. The anthers are removed with small, finepointed forceps, or by suction, or may be killed by heat,
cold, alcohol, or chemical hybridizing agents, and the emasculated flowers are covered with parchment or glassine envelopes, or kraft paper bags, to protect them
from stray pollen (Figs. 12.12, 12.13, and 12.14). Emasculation is unnecessary in monoecious or dioecious crops, although the pistillate flower will need to be
protected from foreign pollen. In highly selfincompatible plants, the greater compatibility of foreign pollen is usually depended upon to fertilize the ovule. Recessive
malesterile genes may be used to eliminate the emasculation procedure in some species. Cytoplasmic male sterility is utilized for production of hybrid seed without
emasculation in onions, corn, pearl millet, sorghum, sugarbeets, sunflower, wheat, and other species.
Pollination Practices
Pollinations are most successful if made within one to three days following emasculation, during the period of maximum stigma receptivity as indicated by the opening
of the flower and full development of the stigma. Pollination is carried out by collecting ripe anthers and emptying the pollen directly upon the stigma, or dusting it on
with a camel hair brush. In small grains and forage grasses, a spike or panicle of the pollen parent can be bagged with the emasculated flower, a procedure called
approach crossing. Shaking the bag daily during the period of stigma receptivity will help to disseminate the pollen. Under normal conditions the length of time that
pollen remains viable varies from one to two minutes for wheat to several hours for corn. Longevity of the pollen is reduced by high temperatures, but may be
extended to several months in some species if dried under vacuum and stored in tight containers at temperatures around 0ºC.
Certain crops, like red clover, alfalfa, and birdsfoot trefoil, are pollinated by insects. In the insect pollinated crops, plants or clones are enclosed in insect proof cages
(Fig. 12.15) and the pollinating insect, commonly bees that have been cleansed of pollen, are introduced into the cage. In these species, a high degree of self
incompatibility is depended upon to prevent self or sibpollination.
Page 233
Fig. 12.12.
Equipment commonly used in emasculating and crossing small grains.
Either curved or straightpointed tweezers may be used. The head
bags are made of parchment or glassine.
Fig. 12.13.
Barley spikes are covered with head bags after
emasculation and pollination. The parents of the
cross are recorded on a tag attached to the barley
spike.
Page 234
Fig. 12.14.
Barley spikes selected to show successive stages in
emasculation and crossing. (A) Spike at the stage of
development when emasculations are normally made in
barley. (B) Spike after emasculation. The glumes have been cut
back to facilitate emasculation and pollination. Note that the
immature florets are closed. (C) Spike at desirable stage
for pollination. The florets are now open. (D) Set of seed
obtained from crossing.
Practices to Enhance Flowering and Seed Production
Crossing is often facilitated by growing parent plants in glass houses, screen houses, or environmentally controlled chambers where temperature, photoperiod, and
humidity favorable for flowering can be maintained and contamination from wind blown pollen reduced. For plants with winter growth habit, vernalization of seeds is
needed for flowering to occur. Vernalization of some species is accomplished by exposing germinating seeds to temperatures slightly above freezing. Success in some
wide species crosses may be enhanced by excising the immature embryo from potentially viable F1 seeds a few days after pollination, and culturing the embryo
aseptically on an artificial medium.
Conducting Field Trials
New cultivars and breeding lines are evaluated in field trials for yield potential, maturity, lodging resistance, resistance to environmental stress, and naturally occurring
disease pathogens
Page 235
Fig. 12.15.
An insectproof cage covered with plastic wire in which experimental strains of legumes are
grown for controlled pollination studies. A hive of honey bees is placed inside the cage to
crosspollinate the legume strains growing in the cage.
or insect pests (Fig. 12.16). Strains with highest performance in preliminary trials are reevaluated in advanced yield trials at different locations for genotype
environment interactions and yield stability. The practice of conducting yield trials for particular crops with a uniform set of entries over the production area was
pioneered by the U.S. Department of Agriculture, and is now widely practiced by the International Research Centers. Additionally, the uniform nurseries serve to
disseminate superior germplasm to the plant breeders who conduct the trials.
Principles in Plot Experimentation
Crop cultivars and breeding lines are grown in field performance trials, simulating performance in natural field environments. The field trial is never conducted in an
ideal environment, some error in the data is universally present. The error may arise from unavoidable or chance fluctuations among the environments in which the
different cultivars are grown, or it may arise from inaccuracy in the conduct of the experiment. If the error is large, the experimenter will not be able to correctly predict
cultivar performance. For accurate and trustworthy results, the experimenter must follow careful and proven procedures, which are uniformly conducted with all of the
entries included in the trial, and personal bias eliminated in recording notes and in interpreting the data.
SOIL VARIABILITY. Variability in the soil is a universal source of error in field plot trials. Even in small contiguous areas, the soil will vary in fertility, drainage,
texture, and productivity; or contain uneven residual effects from previous soil cultural treatments. In selection of the area for field performance trials, consideration
should be given for uniformity in topography, drainage, fertility, and previous cropping treatments.
Page 236
Fig. 12.16.
Wheat breeding nursery conducted by the International Maize and Wheat
Improvement Center (CIMMYT) at Ciudad Obregon, Sonora State, Mexico.
COMPETITION AND BORDER EFFECT. Crop plants in adjacent rows compete for the soil moisture and plant nutrients in the space between them. The error
resulting from competition between adjacent cultivars may be reduced by planting multiplerow plots, and by grouping together cultivars that are similar in maturity,
height, and growth habit. Uniform stands of all cultivars in the trial reduces interplant competition and are essential for accurate yield results.
REPLICATION. In cultivar yield trials, the recorded yield for a cultivar harvested from a single plot is always subject to error; the true yield of the cultivar
(genotype) being either larger or smaller than the recorded yield. If the yields of several plots of the same cultivars are averaged, the chance fluctuations tend to offset
each other and reduce the error. When all entries in an experiment are repeated several times, each repetition is referred to as a replication. The number of
replications in a field experiment normally ranges from three to five, depending upon the design of the experiment, the accuracy desired in the yield data, and the
amount of land and seed available. Replication at a single site is effective in sampling soil variations at that site, and local cultivar × environment interactions. Replication
increases precision in the experiment, accuracy in identifying the superior cultivar in that particular experiment, and provides the means for statistical analysis of the
experiment and estimation of the magnitude of the experimental error.
Page 237
LOCATION AND SEASONAL VARIATION. The comparative performance of crop cultivars differ when yield trials are conducted at a particular location in
different seasons, and when conducted at different locations in the same or different seasons. Each location and\or season provides a different environment in which to
grow the cultivars (genotypes). Replication of an experiment in different seasons and at different locations serves to sample major cultivar × environment interactions
(Fig. 4.1), and assists the breeder in identifying cultivars with stable yield performance over a range of environments. It is also important to examine the cultivar ×
location interactions since cultivars developed with stable yields over a broad range of environments may not be the highest yielding cultivar at a particular location.
Field Plot Design and Techniques
Field experiments are conducted to evaluate comparative cultivar performance and should simulate improved cultural practices as nearly as practical. The design of the
experiment will depend upon the particular crop, the number of cultivars to be tested, and the precision desired in the results. The experimental design must include
replication and randomization of the treatments (cultivars) within the replications in order to make valid comparisons among the cultivars and calculate accurate
estimates of the experimental error. Careful attention to details in planting, harvesting, threshing, and measuring yield are required in order to obtain accurate results.
Cultivar yield trials usually consist of complete or incomplete block designs. The statistical analyses of these designs will not be presented here, but may be found in
standard textbooks on statistical procedures for agricultural research.
COMPLETE BLOCK DESIGNS. The randomized complete block is a simple design in which all treatments (cultivars) are included in each replication of the
experiment and are arranged in a random order within the replication. For accuracy, it should be used with small numbers of cultivars only. Replications may be placed
end to end or opposite each other, so that the total area covered by the experiment will be as nearly square in shape as possible. Entries with apparent weaknesses
may be discarded before harvest, saving the expense of harvest, and the data may still be analyzed by an analysis of variance.
INCOMPLETE BLOCK DESIGNS. With large numbers of entries in yield trials, the size of the experimental area increases, possibly increasing the amount of
error due to soil variation. Incomplete block designs are preferable where large numbers of cultivars are compared in a single yield trial. The cultivars in each
replication are subdivided into smaller blocks, in a manner designed to reduce the error caused by soil variation. The incomplete block designs, usually referred to as
lattice designs, have the restriction that the number of cultivars must be a square of some number. In addition, all cultivars must be harvested; inferior strains cannot be
discarded prior to harvest to reduce harvest expense, and still analyze the experiment as a lattice design.
THE ANALYSIS OF VARIANCE. The analysis of variance is a simple mathematical procedure for measuring the relative importance of two or more groups of
factors that cause variation within an experiment. The analysis of variance is a standard tool for data analysis by plant breeders and quantitative geneticists. Its
applications in experimental design, interpretation of data, and tests of significance are too extensive to be presented here, but may be found in textbooks on statistical
analysis and field plot design.
Page 238
RECORD KEEPING. The plant breeder grows and observes thousands of experimental strains in the breeding nursery, requiring an extensive system of record
keeping. An efficient system of record keeping should possess the following requisites:
• completeness,
• accuracy, and
• simplicity.
Modern computers enable the breeder to summarize, statistically analyze, and reproduce data quickly and accurately. Computerized systems are utilized for
randomization of entries in the yield trials, printouts of notebook sheets and labels for identification of field plots, instant recording of data, and analysis of data.
Study Questions
1. When is emasculation necessary in plant breeding? Describe how the plant breeder goes about making pollinations. What precautions are necessary when making
hand pollinations?
2. How does the plant breeder decide which breeding objectives are important in crop plants?
3. How important are field trials in the overall scheme of plant breeding? What are some of the field experimental designs used in plant breeding?
4. What are some of the techniques used for inducing disease and insect epiphytotics?
Further Reading
Blum, A. 1988. Plant breeding for stress environments. CRC Press, Boca Raton, FL.
Clark, R.B., and R.R. Duncan. 1991. Improvement of plant mineral nutrition through breeding. Field Crops Res. 27:21940.
Clark, R.B., and R.R. Duncan. 1993. Selection of plants to tolerate soil salinity, acidity, and mineral deficiencies. p. 37179. In D.R. Buxton, R. Shibles, R.A.
Forsberg, B.L. Blad, K.H. Asay, G.M. Paulsen, and R.F. Wilson (eds.) International crop science I. Crop Sci. Soc. Am., Inc., Madison, WI.
Clarke, J.M., and T.F. TownieySmith. 1984. Screening and selection techniques for improving drought resistance. p. 13762. In P.B. Vose and S.G. Blixt (eds.)
Crop breeding, a contemporary basis. Pergamon Press, Oxford.
Dyck, P.L., and E.R. Kerber. 1985. Resistance of the racespecific type. p. 469500. In A.P. Roelfs and W.R. Bushnell (eds.) The cereal rusts, Vol. II. Academic
Press, Orlando, FL.
Fehr, W.R., and H.H. Hadley. 1980. Hybridization of crop plants. Am. Soc. Agron., Crop Sci. Soc. Am., and Soil Sci. Soc. Am., Madison, WI.
Fowler, D.B., A.E. Limin, A.J. Robertson, and L.V. Gusta. 1993. Breeding for lowtemperature tolerance in field crops. p. 35762. In D.R. Buxton, R. Shibles,
R.A. Forsberg, B.L. Blad, K.H. Asay, G.M. Paulsen, and R.F. Wilson (eds.) International crop science I. Crop Sci. Soc. Am., Inc., Madison, WI.
Gomez, K.A., and A.A. Gomez. 1976. Statistical procedures for agriculture research. The Int. Rice Res. Inst., Los Baños, Philippines.
Harpstead, D.D. 1983. Breeding for improved nutritional quality of crops. p. 25570. In D.R. Wood (ed.) Crop breeding. Amer. Soc. Agron., Crop Sci. Soc. Am.,
and Soil Sci. Soc. Am., Madison, WI.
Hooker, A.L. 1983. Breeding to control pests. p. 199230. In D.R. Wood (ed.) Crop breeding. Am.
Page 239
Soc. Agron., Crop Sci. Soc. Am., and Soil Sci. Soc. Am., Madison, WI.
Johnson, R., and G.J. Jellis (eds.) 1993. Breeding for disease resistance. Kluwer Academic Publishers, Dordrecht, The Netherlands.
Parlevliet, J.E. 1985. Resistance of the nonrace specific type. p. 50125. In A.P. Roelfs and W.R. Bushnell (eds.) The cereal rusts, Vol. II. Academic Press,
Orlando, FL.
Pinthus, M.J. 1973. Lodging in wheat, barley, and oats: The phenomenon, its causes, and preventive measures. Adv. Agron. 25:20963.
Rasmusson, D.C., and B.G. Gengenbach. 1983. Breeding for physiologic traits. p. 23154. In D.R. Wood (ed.) Crop breeding. Am. Soc. Agron., Crop Sci. Soc.
Am., and Soil Sci. Soc. Am., Madison, WI.
Rowell, J.B. 1984. Controlled infection by Puccinia graminis f. sp. tritici under artificial conditions. p. 291332. In A.P. Roelfs and W.R. Bushnell (eds.) The cereal
rusts, Vol. I. Academic Press, Orlando, FL.
Sleper, D.A., T.C. Barker, and P.J. BramelCox (eds.). 1991. Plant breeding and sustainable agriculture: Considerations for objectives and methods. CSSA special
publ. No. 18. Am. Soc. Agron., Crop Sci. Soc. Am., and Soil Sci. Soc. Am., Madison, WI.
Smith, C.M., Z.R. Khan, and M.D. Pathak. 1993. Techniques for evaluating insect resistance in crop plants. Lewis Publishers, Boca Raton, FL.
Shedecor, G.W., and W.G. Cochran. 1991. Statistical methods. 8th ed. (2nd Printing). Iowa State University Press, Ames, IA.
Sojka, R.E. 1985. Field evaluation of drought response in smallgrain cereals. p. 16591. In G.E. Russell (ed.) Progress in plant breedingI. Butterworths, London.