Chapter 4
Breeding Schemes

4.1 Introduction

All successful breeding programmes have been designed around a breeding scheme. The breeding scheme determines the passage of breeding lines through the selection process, and through to the increase of planting material for cultivar release. The process of selection will be carried out over a number of years, and under differing environmental conditions. The early selection stages of breeding programmes will involve screening many thousands of different genotypes. The early screening is therefore relatively crude, and in many instances this involves only visual selection. After each round of selection, the ‘better’, more adapted, or more disease-resistant genotypes will be retained for further evaluation while the least adapted lines will be discarded. This process will be repeated over a number of years, and at each stage the number of individual genotypes or populations is reduced and evaluation is conducted with greater precision in estimating the worth of each entry.

The breeding scheme used will be highly dependent upon the crop species and the type of cultivar (pure-line/self-pollinated, open-pollinated, hybrid, clonally propagated, synthetic, etc.) that is being developed. The general philosophy for developing a clonal cultivar like potato is therefore different from a pure-line cereal cultivar, say barley. In the former, breeding selections are genetically fixed through vegetative propagation, but there will be a low rate of multiplication of planting material. In the latter, there will be more rapid increase of planting material, although the segregating nature of the early-generation breeding lines will complicate the selection process.

The most effective breeding schemes will utilize the positive attributes of a crop species while minimizing difficulties that might arise through the selection process. In the following section the general breeding schemes for pure-line, open-pollinated, hybrid and clonally propagated cultivars will be explained, along with mention of the schemes used for developing multilines and synthetics.

4.2 Development of pure-line cultivars

Crops that are generally produced as pure-line cultivars include barley, chickpea, flax, lentil, millet, peas, soybean, tobacco, tomato and wheat.

One and a half centuries ago, most inbred crop species were grown in agriculture as ‘landraces’. Landraces were locally grown populations which were, in fact, a collection of many different genotypes grown in mixture and which were, of course, both genetically and phenotypically variable. Pure-line (synonymous with inbred) cultivars were developed first from these landraces by farmers who selected specific (presumably more productive or least infected with disease or pests) plants from the mixed populations and maintained these in isolation, thus encouraging selfed progenies, and eventually developed homozygous, or near-homozygous, lines. It is reasonable to assume that these homozygous lines were indeed more productive than the original landraces because by the end of the 19th century, landraces had almost completely disappeared in countries with advanced agricultural systems.

These early ‘pure-line breeders’ used the naturally existing genetic variation within the landraces they were propagating and the natural tendency of some species to self-pollinate (e.g. wheat) at a high frequency. However, this strategy has a limited potential in terms of generating new variation, and so modern plant breeders have to continuously generate genetic variation and hence the three-phase breeding schemes were established to create genetic variation, identify desirable recombinant lines within progenies and stabilize and increase the desired genotype. It is interesting to note, however, that recently a number of plant breeders have returned to old landraces of wheat and barley to examine their wealth of genetic diversity, as well as to testing combinations of lines in ‘modern’ landrace combinations (i.e. multilines). Unfortunately most of the landraces that existed even 100 years ago are no longer available and potentially valuable germplasm and adapted combinations have been lost.

By far the most commonly used method of generating genetic variation within inbreeding species is via sexual reproduction using artificial hybridization. There are, of course, other ways to produce genetic variation. For example, variation can be produced by induced mutation, somatic variation, somatic hybridization and recombinant DNA techniques (all discussed in later chapters).

After sexual crossing between genotypes and then selfing them to generate suitable genetic variation, plant breeders will then traditionally screen the segregating population for desirable ‘segregants’ while continuing to self-pollinate successive generations, to produce homozygous lines. Thus, accomplishing the last two steps of the breeding scheme (selection and stabilization) can be achieved more or less simultaneously.

4.2.1 Homozygosity

One of the difficulties in selecting desirable recombinant lines in pure-line breeding is related to segregating populations and the masking of specific character expression as a result of the dominant/recessive nature of the segregating alleles in the heterozygotes. Another consideration is the relationship between genetic homozygosity and ‘commercial inbred lines’. The definition of complete homozygosity is that all the alleles at all loci are identical by descent, that is, there is no heterozygosity at any locus. However, for practical commercial exploitation, the level of homozygosity does not need to be complete. Clearly the lines must basically breed ‘true to type’ but this is by no means absolute. The degree of homozygosity is determined by the level of inbreeding and directly related, for example, to the number of selfing generations that have been performed. Consider the simple case of just one locus with two possible alleles A-a:

Consider now a more complex situation where six loci are involved, as set out below. Of these six loci, two, loci and , have the same allele in both parents (and are both homozygous) and so the is also homozygous at these two loci. At the other four loci the parents are homozygous but have different alleles, and so the is heterozygous at these loci and subsequently these segregate in the .

Parents
	AABbCcDdEeff
	AABBCcDdeeff, AABbCcDDEeff,
	AAbbCCEeDdff, AABBccDdEeff, etc.

This can be generalized in mathematical terms as follows. Consider an that is heterozygous at loci; heterozygosity at any single loci after generations of selfing will be:

Therefore the probability of homozygosity at all loci will be:

Hence after generations:

This can also be written as:

The level of homogeneity (i.e. uniformity in appearance/phenotype as opposed to genetic homozygosity) required in an inbred cultivar will depend to a varying extent on the personal choice of the breeder, seed regulatory agencies, farmers and end-users. Almost all pure-line breeding schemes involve selection of individual plants at one or more stages in the breeding scheme. The stages of single plant selection will have a large impact regarding the degree of heterogeneity in the end cultivar. If single plant selection is carried out at an early generation, say there may be greater heterogeneity within the resulting cultivar compared with a situation where single plant selection was delayed until a later generation, say when individual plants would be more homozygous. Breeders must ensure that a level of uniformity and stability exists throughout multiplication and into commercialization. Farmers and processors (such as millers) will have preferences for cultivars that are homozygous, and hence homogeneous, for particular characters. These characters may be related to uniform maturity, plant height or other traits related to ease of harvest. Many believe that farmers are not concerned with uniformity of characters that do not interfere with end-use performance (e.g. flower colour segregation). However, farmers take a natural pride in their farms and, therefore, like to grow ‘nice-looking’ crops, and these are ones which are uniform for almost all visible characters. In any case they certainly like uniformity in terms of characters such as harvest date! For most end-users there will be an obvious preference for cultivars with high uniformity of desirable quality characters. For example, there may be a premium for uniform germination in malting barley or more uniform characters relating to breadmaking in wheat.

In contrast, some breeders of pure-line cultivars like to maintain a relatively high degree of heterogeneity in their developed cultivars. They believe that this heterogeneity can help to ‘buffer’ the cultivar against changes in environment and hence make the cultivar more stable over different environments. Often statutory authorities determine the degree of variability that is allowed in a cultivar. For example, all pure-line cultivars released in the European Union countries, along with Canada and Australia, must comply to set standards for distinctness, uniformity and stability (DUS) in Statutory National Variety Trials. In these cases it is common to have almost total homogeneity and homozygosity in released inbred cultivars.

For most breeders, time is at a premium. Therefore, some methods are commonly used to reduce the time taken to achieve homozygosity, and these include single seed descent and the use of off-season sites (this excludes the production of homozygous lines through doubled haploids, which is relevant here but will be discussed separately in a later chapter).

Single seed descent

Single seed descent involves repeatedly growing a number of individuals initiating from a segregating population, usually under high-density, low-fertility conditions, to accelerate seed-to-seed time. At maturity, a single seed from the natural self of every plant is taken and replanted. This operation is repeated a number of times to obtain homozygous plants. Single seed descent is most suited for rapid generation increase in a greenhouse where a number of growth cycles may be possible each year. Single seed descent in canola, wheat and barley can be further accelerated by growing plants under stress conditions of high density, high light, restricted root growth, and low nutrient levels, which result in stunted plants with only one or two seeds each, but in a shortened growing period compared with growth under normal conditions (up to three or four generations in a year are possible in barley or spring canola).

It is very important, however, when using single seed descent, that unintentional selection is not being carried out for adverse characters. For example, in a single seed descent scheme in winter wheat (where plants will require a vernalization period prior to initiating a reproductive phase), vernalization requirements may be overcome artificially in a cold room. If this is done, then care should be taken so that all seedlings do indeed receive sufficient cold treatment to overcome the vernalization requirement – otherwise the system will select the plant types with lower vernalization requirements. In addition, some genetic characteristics are not fully expressed when plants are grown under high competition stress conditions used for single seed descent. For example, the erectoides dwarfing gene (ert) in barley plants is not expressed under single seed descent in the glasshouse, and therefore genotypes cannot be selected for dwarfism under these conditions. In any case it is strongly advised that no selection be practised during this phase.

Off-season sites

Off-season growing sites can also reduce the time for achieving a desired level of homozygosity. This is possible by having more than one growing season per year. Dual locations at similar latitudes in the northern and southern hemispheres are frequently used to increase either seed quantity or reduce heterozygosity in many breeding schemes. The use of off-season sites is often restricted to annual spring crops, and there are only a few good examples where they have helped accelerate homozygosity in winter annuals, and virtually none in breeding biennials.

If off-season sites are to be incorporated into the breeding scheme, care must be taken to ensure that ‘selectional adaptability’ of the off-season site does not have adverse effects on the segregating plant populations. For example, the spring barley breeding scheme at the Scottish Plant Breeding Station used to increase breeding selections to by growing these lines over winter in New Zealand. Although New Zealand has a climate that is very similar to that found in Scotland, there is a completely different spectrum of races of powdery mildew. As a result, mildew-resistant selections made in New Zealand were of no relevance when grown in Scotland, and so meant that all New Zealand trials needed careful spraying to avoid powdery mildew being confounded with other performance characters. Breeding companies developing varieties for the US market have developed large networks of off-season sites around the globe for crops such as maize and soybean.

The use of off-season sites benefits farmers as they significantly reduce the time required to take new varieties to the market, and so to respond to the ever-changing needs of humankind.

4.2.2 Breeding schemes for pure-line cultivars

There are probably as many different breeding schemes used by breeders of pure-line cultivars as there are breeders of inbreeding species. There are, however, three basic schemes: bulk methods, pedigree methods and bulk/pedigree methods. It should be noted that all the breeding schemes described involve more than a single cross at the crossing stage. A number of these crosses will be two-parent crosses (female parent, say although many breeders use three- and four-way parent cross combinations ( and respectively).

Bulk method

The outline of a bulk scheme is illustrated in Figure 4.1. In this scheme, genetic variation is created by artificial hybridization between chosen parents.

Figure 4.1 Outline of a bulk breeding scheme used for breeding pure-line crop species.

The and several subsequent generations, in the illustration up to and including the generations, are grown as bulk populations. These bulks are left to set seed naturally which, with these species, will mean predominantly self-pollination. No conscious selection is imposed in these generation, and it is assumed that the genotypes most suited to the environment in which the bulk populations are grown will leave more offspring and hence predominate in future generations. Similarly, these bulk populations are usually grown under the stress and disease pressures common to the cultivated crop, and it is assumed that the frequency of adapted genotypes in the population increases. It is therefore very important that the bulks are grown in a suitable and representative environment. After a number of rounds of bulk increase, individual plants showing desirable characteristics are selected, often at the stage. From each selected plant, a head of seed is taken and grown as a row (a head-row). The produce from the best head-rows are bulk harvested, for initial yield trials. More advanced yield trials are grown from bulk harvest of desirable individuals.

The major advantage of the bulk method is that conscious selection is not attempted until plants have been selfed for a number of generations and hence the plants are nearly homozygous. This avoids the difficulty of selection among segregating populations where phenotypic expression will be greatly affected by levels of dominance in the heterozygotes. This method is also one of the least expensive methods of producing populations of inbred lines. The disadvantage of this scheme is the relatively long time from initial crossing until yield trials are grown. In addition, it has often been found that the natural selection, which occurs through bulk population growth, is not always that which is favourable for growth in agricultural practice. In addition, natural selection can, of course, only be effective in environments where the character is expressed. This often prevents the use of bulk methods at off-season sites.

Other methods have been used to produce homozygosity in bulk breeding schemes. These include single seed descent and doubled haploidy. Breeding schemes that use these techniques have increased the popularity of bulk breeding scheme in recent years, as the time from crossing to evaluation can be minimized. However, the basic philosophy is similar, being to produce near-homozygous lines, and thereafter select amongst these. Where rapid acceleration to homozygosity techniques are used, it is essential to ensure that no negative selection occurs. For example, research has shown that creating homozygous breeding lines of canola (B. napus) through pollen culture produces a higher than random frequency of plants with low erucic acid in the seed oil. If low erucic acid content is desired, this poses a selection advantage. If, however, an industrial oilseed cultivar were desired (one with high erucic acid content), then using embryogenesis would be detrimental.

Pedigree method

The outline of a pedigree breeding scheme is shown in Figure 4.2. In a pedigree breeding scheme, single plant selection is carried out at the through to generations. The scheme begins by hybridization between chosen homozygous parental lines, and segregating families are obtained by selfing the heterozygous s. Single plants are selected from amongst the segregating families. The produce from these selected plants is grown in plant/head rows at the generation. A number of the ‘most desirable’ single plants (in Figure 4.2, four plants) are selected from the ‘better’ plant rows and these are grown in plant rows again at the stage. This process of single plant/head selection is repeated until plants are ‘near’ homozygous (i.e. .) At this stage the most productive rows are bulk harvested and used as the seed source for initial yield trials at .

Figure 4.2 Outline of a pedigree breeding scheme used for breeding pure-line crop species.

In addition to being laborious (as a considerable amount of record-keeping is required) and relatively expensive, annual discarding may lead to the loss of valuable genotypes, particularly under the changing environmental conditions from year to year, making selection difficult. Other disadvantages of the pedigree method are that it requires more land and labour than other methods; experienced staff with a ‘good breeders’ eye' are necessary to make plant selections; selection is carried out on single plants where errors of observation may be great; while actual yield testing is not possible in the early generations.

If selection was effective on a single plant basis, pedigree breeding schemes would allow inferior genotypes to be discarded early in the breeding scheme, without the need for being tested in more extensive, and costly, yield trials. Unfortunately, pedigree breeding schemes offer little opportunity to select for quantitatively inherited characters, and even single gene traits can cause problems when selecting on a single plant basis in highly heterozygous populations.

Bulk/pedigree method

The outline of a bulk/pedigree breeding scheme is illustrated in Figure 4.3. This type of breeding scheme uses a combination of bulk population and single plant selection. An population is produced by selfing plants from controlled hybridizations. Individual plant selections from the segregating families are grown in plant progeny rows at the stage. Selected families are bulk harvested and preliminary yield trials are grown at the stage by planting the bulked seed. and bulk seed yield trials are grown, in each case by planting bulked seed from the previous year's trial. Selection of families is based upon performance in these trials. At the stage, single plant selections are once again made from the now near-homozygous lines. Progeny from these plant selections are grown then as plant rows at ; second-cycle initial yield trials at ; and more advanced yield trials at

Figure 4.3 Outline of a bulk/pedigree breeding scheme used for breeding pure-line crop species.

The advantage of this combined breeding scheme is that inferior individuals, lines, families or populations are identified and discarded early in the breeding scheme. More than a single cultivar may be derived from a family or heterogeneous line identified as being superior by the earlier generation testing. Disadvantages will include, with fixed resources, the use of testing facilities for evaluation of individual lines in the early generations and so reducing the number of the later, more highly inbred lines that can be evaluated. Despite these disadvantages, bulk/pedigree schemes (or close derivatives of) are most commonly used to develop pure-line cultivars.

Modified pedigree method

Most breeding schemes have developed breeding schemes that are combinations of bulk and pedigree methods. For example, the breeding scheme used for developing winter barley cultivars at the Scottish Plant Breeding Station was a modified pedigree breeding scheme (illustrated in Figure 4.4).

Figure 4.4 Outline of a modified pedigree breeding scheme used for breeding pure-line crop species.

The modified pedigree breeding scheme enables yield trials to be grown simultaneously with pedigree selection. Single plants are selected from amongst segregating families. Seed from these selections are grown as plant progeny rows at . One, or more, single plants are selected from each of the desirable plant rows, and the remainder of the row is bulk harvested. The single plant selections are grown as plant progeny rows at It is common in many crops, for example wheat and barley, to select single ears/heads rather than whole plants. In this case, the plant progeny rows are commonly referred to as ‘head-rows’ or ‘ear-rows’, as seed from a single ear/head is used to plant a single row. Simultaneously, the harvested bulk is planted in a preliminary yield trial. The seed from the bulk yield trial is used to plant a more extensive bulk yield evaluation trial at the stage. Based on the results from the bulk yield trial, the most productive populations are identified. Single plant selections are made from the corresponding plant progeny rows and the remaining row is bulk harvested for a further yield trial the following year at the generation. This process is repeated at the to stage, by which time near homozygosity is achieved in the remaining lines.

The advantage of the modified pedigree breeding scheme is that it utilizes progeny bulk evaluation for yield and other quantitatively inherited characters, while single gene traits can be screened on a plant progeny row basis. In addition, this scheme allows for evaluation of quantitative characters while simultaneously inbreeding the selections.

4.2.3 Number of segregating families and selections

There have been numerous debates amongst plant breeders concerning the question of how many plants or families should be evaluated, and selected, at each stage in a breeding scheme. Unfortunately, there is no simple recipe to help new breeders, and the questions can only be addressed from an empirical standpoint. Plant breeding is a numbers game and the chance of success is often associated with screening many thousands of breeding lines. However, plant breeding programmes should only be as large as the specific breeding group can handle and as resource availability allows. Therefore, it is not productive to grow more lines at any stage in a breeding scheme than can be effectively and accurately assessed.

It is often easier to work backwards and ask how many lines can be handled at, say, the advanced yield trial stage in the breeding scheme, and then move backwards to the previous stage and predict how many lines are required at that stage to ensure that the required number are selected, and so on.

Similarly, the number of initial cross combinations that should be used differs markedly in different breeding programmes. Often a large number of crosses need to be screened, as the breeder cannot identify the most productive cross combinations. With experience of specific parents in cross combination and the benefit of ‘cross prediction’ techniques (see Chapter 7), it is possible to reduce the number of crosses screened on a large scale and hence allow breeders to put greater emphasis on cross combinations with the highest probability of producing desirable recombinants and hence cultivars.

4.2.4 Seed increases for cultivar release

At the other end of the breeding programme, once desired cultivars have been identified, it is necessary to produce a suitable quantity of seed that will be grown and increased for varietal release. This ‘seed lot’ is usually called Breeders' seed, as in most cases producing this seed is the responsibility of the breeder rather than a ‘seedsman’. It is vital that breeders' seed lots are pure, free from variants, and that the cultivar that is to be released is ‘true-to-type’. Breeders' seed is used in multiplication to produce ‘foundation’ seed, which, in turn, is used to produce the various levels of ‘registered’ or ‘certified seed’, which is eventually sold to farmers.

Producing high quality breeders' seed is very similar to the breeding schemes (described above). In general there are two basic methods of producing breeders' seed, mass bulk increase and progeny test increase.

Mass bulk increase

In mass bulk increase schemes, a uniform sample of seed from the selected line is chosen and planted on only one occasion to result in the breeders' seed lot. The advantage of this simple method is that it is inexpensive and takes only a single season to obtain the required seed. The disadvantages are mainly related to the purity, homozygosity and homogeneity of the cultivar entering into commercialization.

Progeny test increase

The progeny test increase method is more expensive and takes longer to obtain the seed required. This method is very similar to the bulk/pedigree breeding scheme. A number of single plants are selected from the homozygous/near homozygous advanced breeding line. These are grown as plant progeny rows. Individual plant rows are discarded if they display off-types or are non-uniform. The remaining rows are harvested individually and the seed from each row is used to plant larger progeny plots the following season. These plots are again inspected and those that do not have the required homogeneity or show off-types are discarded. At harvest the progeny plots are bulk harvested as breeders' seed. Breeders who wish to maintain a degree of heterozygosity in the released cultivar will include greater numbers of initial single plant selections in the scheme, or they may not be as restrictive in decisions to discard progeny rows or plots. The advantage of the progeny multiplication method is that it allows greater control by the breeder and results in greater homogeneity in the released cultivar.

4.3 Developing multiline cultivars

Multiline cultivars (multilines) are mixtures of a number of different genetic lines, often varying in their disease resistance genes. Multiline cultivars are almost exclusively composed of mixtures of lines from pure-line species. Multilines have been developed for a number of different crop species including barley, wheat, oats and peanuts. In turf grass, intraspecific and interspecific multiline cultivars are grown commercially.

Multilines have been suggested as one means to minimize yield or quality losses due to diseases or pests that have multiple races and where the race specificities can change from year to year. In this latter case there is a lower probability that all plants within the multiline (with a range of specific disease resistance genes or resistance mechanisms), as opposed to a pure-line cultivar, would be affected as severely as when the population carries only one resistance gene providing protection to just one pathogen race. It has also been suggested that the use of multilines would result in more durable mechanisms of disease resistance in crop species.

Research has also suggested that multiline cultivars are more stable over a range of different environments than are pure-line cultivars. The reason for this has been related to the heterogeneous nature of the mixture where some lines in the mix do well in some years or locations while others perform better under other conditions. Therefore multilines show fewer genotype by environment interactions, a primary reason for their popularity with peanut breeders. Similarly, such considerations have led to mixtures of rye grass and Kentucky bluegrass being sold commercially. Rye grass has rapid emergence and establishment and does better than bluegrass in shaded areas.

Marketing of multiline varieties in the US has advantages over other types of cultivar as seed companies can sell the seed without a common brand name if the seed sold is labelled with a ‘cultivar not stated’ label. Multilines can also be sold under more than one name. For example, the same multiline can be sold with the brand names ‘Browns Appeal’, ‘Browns Wonder’ and ‘Wonder Why’ by the same or different seed sales groups. In other countries, however, multiline cultivars must comply with the set standards of DUS (Distinctiveness, Uniformity and Stability) required for other inbred cultivars, and this has limited their use because of the difficulties in obtaining such homogeneity standards in a mixture.

The same care needs to be taken when producing breeders' seed for a multiline cultivar as is the case with a pure-line cultivar. The individual lines forming the mixture are increased independently by either mass or progeny multiplication methods (above). The individual components are then mixed in the proportions required, the seed mixed to form the breeders' seed, from which foundation seed is produced. The prevalent diseases, yielding ability or other appropriate factors will determine the proportion of lines within the mix. It is important when calculating multiline mixture proportions to take into account the seed size (if mixing by weight) and also the germination potential of each line (which may be different for the different lines).

One major complication relating to seed mixture proportions is the reproductive potential or productivity of each genotype in the mixture. For example, if the given proportion of two-parent lines (A and B) in a very simple multiline is 1:1, but the reproductive potential of A is twice that of B, a 1:1 mix of breeders' seed will result in a 2:1 ratio of the lines being harvested from foundation seed and a 4:1 in registered seed, and finally a 8:1 ratio being sold to the farmer after one further round of multiplication, to obtain certified seed. Similarly, environmental conditions may affect the proportions of mixed lines in the multiline. These changes could reflect that foundation and certified seed were produced in an atypical environment or where a different disease spectrum exists.

Some multiline cultivars are mixtures of isogenic (or near-isogenic) lines that differ for a single gene (usually conferring resistance to a certain strain of a pathogen). The most common method used in developing isogenic lines (lines that only differ in their genotypes by specified genes) in plant breeding is through the use of backcrossing. The genetics of backcrossing will be covered later in the qualitative genetics section. However, a brief description will be given here.

4.3.1 Backcrossing

Backcrossing is a commonly used technique in developing pure-line cultivars. This technique has been used in plant breeding (not only in inbred species) to transfer a small, valuable portion (often ideally a single gene) of the genome from a wild or unadapted genotype into the background genotype of an adapted and already improved cultivar.

Backcrossing is an operation that involves a recurrent parent and a non-recurrent parent. In many cases the non-recurrent parent is an unadapted line or genotype, and hence is not expected to contribute characters other than the specific one that it is desired to transfer to the resulting selections. It is therefore usual to choose a recurrent parent which is already suited to the environment (i.e. the most adapted genotype or cultivar available). The process of isogenic line production will be identical and repeated for each line that is to be produced.

The process will be described for the case where the homozygous allele (RR) of interest from the non-recurrent parent is completely dominant in showing resistance to a disease and the recurrent parent has the recessive susceptible alleles (rr). First the recurrent parent is crossed to the non-recurrent parent producing seeds which are therefore heterozygous (Rr) for this locus and where each of the two parents contribute equally to the genotype. The s (Rr) are crossed back to the recurrent parent , to produce backcross 1 . The seeds from this ‘backcross’ are, for this locus of the genotypes Rr or rr, in equal proportions, which can be screened to identify the disease-susceptible lines rr, as opposed to the disease-resistant Rr. The Rr lines can then be used to cross back again to the recurrent parent rr, to produce the second backcross or which again comprises the same genotypes as the . This process of screening for the presence of the heterozygous resistant lines and backcrossing them to the recurrent parent is repeated a number of times with the aim of developing a line which comprises all the genes from the recurrent parent except at the ‘resistance locus’, which will have the resistance allele instead of the susceptible allele . In other words we effectively ‘add’ this to the genotype. The number of backcrossing generations will depend on how closely the breeder wants the isogenic line to resemble the recurrent parent or how well the backcross genotypes are performing. The proportion of the recurrent parent genotype in each backcross family will increase with increased backcrossing, and can be calculated by the formula:

where is the number of backcrossing generations, including the original cross to produce the . The following are proportions of the genes that are theoretically recovered from the recurrent parent according to the number of backcrosses:

The above percentages of the recurrent parent genotype in the resulting progeny hold reasonably well in the early backcross generations. However, with increased backcrossing, the percentage of the genotype of the non-recurrent parent (the ‘wild’ type) that is retained will be increasingly influenced by linkage. The resulting backcross line can therefore often contain a higher proportion of wild-type genotype than desired, and more backcross generations will be required to obtain the desired proportion of the cultivated/adapted (recurrent) parent.

Once the desired number of backcrossing generations has been completed, then the heterozygous (Rr) disease-resistant lines are selected, self-pollinated or inter-mated, and lines homozygous for the resistance gene (RR) can be identified.

The backcrossing method where the gene of interest is recessive is a slightly more complex (and often more lengthy) process. The general theory is the same but in this case it is necessary to progeny test the backcrossed generations in order to separate the homozygous and heterozygous plants that need to be selected. Progeny testing can be avoided (or reduced) when tightly linked co-dominant molecular markers are available. Molecular markers also can be used to increase the frequency of the adapted (recombinant) parent genome in the backcross family. This is discussed in more detail in Chapter 8.

4.4 Development of open-pollinated population cultivars

Crops that are generally produced as open-pollinated population include alfalfa, forage legumes, herbage grasses, maize (some), oil palm, perennial ryegrass, red clover, rye and sugar beet.

Development of open-pollinated population cultivars is a process that changes the gene frequency of desirable alleles within a population of mixed genotypes while trying to maintain a high degree of heterozygosity. So it is really the properties of the population that are vital, not individual genotypes (as in pure-line cultivars of self-pollinating crops). Instead of ending with a cultivar for release that is a uniform genotype, the population will be a complex mixture of genotypes, which together give the desired performance.

It is not considered desirable (and it is often very difficult) to develop homozygous or near-homozygous breeding lines or cultivars from these outbreeding species as they suffer severe inbreeding depression, carry deleterious recessive alleles or have strong or partial self-incompatibility systems. There are basically two different types of outbreeding cultivars available: open-pollinating populations and synthetic cultivars.

4.4.1 Breeding schemes for open-pollinating population cultivars

In open-pollinating populations, selection of desirable cultivars is usually carried out by mass selection, recurrent phenotypic selection or selection with progeny testing. Open-pollinating (outbred or cross-pollinating) cultivars are maintained through open-pollinated populations resulting from random mating.

Mass selection

Mass selection is a very simple breeding scheme that uses natural environmental conditions to alter the allelic frequencies of an open-pollinating population. A new population is created by cross-pollinating two different existing open-pollinating populations. In this case a representative set of individuals from each population will be taken to be crossed – single plants will not of course be representative of the populations. So it is common (even if mistaken) to select individual plants from each population but to take a reasonably large sample of such plants. How they are crossed depends upon the choice of the breeder, but often Bi-Parental matings (BIPs) are performed where specific parents are selected and hybridized. Alternatively, it is common to collect pollen in bulk and use this to pollinate specifically selected female plants in another population, whilst in many other instances breeders allow random mating or open/cross pollination to occur naturally.

The seed that results from such a set of crosses is grown under field conditions over a number of seasons. The theory of the approach is that genotypes that are adapted to the conditions will predominate and be more productive than those that are ‘less fit’. It is also assumed that crossing will be basically at random and result in a population moving towards equilibrium.

Problems with mass selection are related to partial, or complete lack of, control of the environmental conditions other than by choosing suitable locations for the trials. In some instances it is possible to create disease stress by artificially inoculating susceptible plants with a pathogen to act as spreaders, or by growing very susceptible lines in close proximity to the bulk populations. However, the process is empirical and often subject to unexpected disturbances. It also assumes, as noted earlier, that natural selection is going to be in the direction that the breeder desires – an assumption that is not always justified. It should also be noted that care needs to be exercised in isolating this developing population from other crops of this species, which might happen to be growing within pollination distance.

Recurrent phenotypic selection

Recurrent phenotypic selection tends to be more effective than mass selection. The basic outline of this process is illustrated in Figure 4.5. A population is created by cross-pollination between two (or more) populations to create what is referred to as the base population. A large number of plants are grown from the base population and a subsample of the most desirable phenotypes is identified and harvested as individual plants. The seeds of these selected plants are grown out, allowed to cross-pollinate at random, and thus to produce seed for a new population – an improved population.

Figure 4.5 Basic recurrent phenotypic selection scheme.

This process is repeated a number of times, in other words, it is recurrent. The number of cycles performed will be determined by the desired level of improvement required over the base population, the initial allele frequencies of the base population, and the heritability of the traits of interest in the selection process. Recurrent phenotypic selection has been shown to be effective, but mainly in cases where there is high heritability of the characters being selected for, such as disease and pest resistances. The techniques are not nearly as effective where traits have a lower heritability, such as yield or quality traits.

It is common practice (and a good idea) to retain a sample of the base population so that the genetic changes due to selection can be evaluated in a later season.

Progeny testing

There are actually a variety of possibilities within the main heading of recurrent phenotypic selection, for example half- or full-sib selection with test crossing; and selection from progeny testing. All the schemes basically involve selecting individuals from within the population, and crossing or selfing these to produce seed. Part of the seed is sown for assessment and part is retained. Once the results of the assessments are available, the remnant seed from the progenies that have been shown to be superior are then sown as a composite population for plants again to be selected, and so on. At any stage seeds can be taken out for commercial exploitation.

4.4.2 Backcrossing in open-pollinated population cultivar development

Backcrossing is not as commonly used with open-pollinated population cultivar development, unlike pure-line cultivar development, but nevertheless the backcrossing technique is still used. When used, the main difference is that the recurrent and non-recurrent parents in the backcrosses are plant populations rather than homozygous lines. The basic assumption of any backcrossing system is that the technique is unlikely to result in a change in performance of the recurrent parent, other than for the single character being introduced. However, even when an allele has been introduced it is difficult to ensure its even distribution throughout the whole population.

Seed production

In most cases, seed production of open-pollinated cultivars simply involves taking a sample of seed from the population but under increasingly stringent conditions, to avoid the problems noted before, in avoiding contamination or cross-pollination from other populations or cultivars.

4.5 Developing synthetic cultivars

A synthetic cultivar basically gives rise to the same end result as an open-pollinated cultivar (i.e. population improvement), although a synthetic cultivar cannot be propagated by open-pollination without changing the genetic make-up of the population. This has perhaps been a primary reason for the rapid change-over from open-pollinated cultivars to synthetic ones, since it means that farmers need to return to the seed companies for new seed each year. It has been commercial seed companies that have been responsible for breeding almost all synthetic varieties. For example, before 1950 in the USA there were only two alfalfa breeders working in private seed companies while 23 were breeding in the public sector. By 1980 there were 17 public-sector alfalfa breeders, but now there are more than three times (52) the number of private breeders developing synthetic lines.

A synthetic cultivar must be reconstructed from its parental lines or clones. Within the US, maize is almost exclusively grown as hybrid cultivars whereas in many other countries maize crops are grown as synthetics. Synthetics have also been used almost exclusively in the development of alfalfa, forage grass and forage legume cultivars, and have also been used to develop varieties from other crop species (e.g. canola).

The breeding method used for the development of synthetic cultivars is dependent on the ability to develop either homozygous lines from a species or to propagate parental lines clonally. In the case of maize, for example, synthetic cultivars are developed using inbred lines as a three-stage process:

• develop a number of inbred lines;
• progeny-test the inbred lines for general combining ability;
• identify the ‘best’ parents and intercross these to produce the synthetic cultivar.

This process is almost identical to the procedure used for developing hybrid cultivars, and only differs in the last stage where many more parents will be included in a synthetic than in a hybrid cultivar. To avoid repetition, therefore, this section on synthetics will only cover the case of developing synthetic cultivars where inbred line development is not possible (for example, due to high inbreeding depression. So we will deal here with clonal synthetics. The process of developing a synthetic cultivar from clonal lines is illustrated in Figure 4.6.

Figure 4.6 Breeding procedure used to develop synthetic cultivars from a clonally reproduced open-pollinated crop species (i.e. alfalfa) and using polycross progeny testing.

Following Figure 4.6, clonal selections for use as parents can be added to the nursery ad infinitum, on the basis of continued phenotypic recurrent selection of the base population or by selections from those that have been produced by designed cross-pollinations. The second stage involves clonal evaluation and is conducted using replicated field trials of asexually reproduced plant units. The aim of the clonal testing is to identify which clonal populations are phenotypically most suited to the environments where they are to be grown. The clonal trials are often grown in two or more locations to include an assessment of environmental stability

A test cross or polycross technique will then be used to ‘genetically’ test the ‘best’ clones identified from the clonal screen. The aim of this ‘genetic’ test is to determine the general combining ability of each clonal line in cross-combinations with other genotypes in the selected group of clones.

If a test cross (often called a ‘top cross’) is used, all the selected clones are hybridized to one (or more) test parent. The test parent will have been chosen because it is a desirable cultivar or it may be chosen because of past experience of the individual breeder. The test parent is a heterozygous clone that produces gametes of diverse genotypes. This diversity of gametes produced from the tester will help an assessment of the average ability of each clone to produce superior progeny when combined with alleles from many different individuals. Test cross evaluations are most useful when the variation that is observed within the different progeny is a result of differences between clones under evaluation and not due to only a small sample of genes coming from the test parent.

A polycross does not use a common test parent but rather a number of different parents. It therefore differs from a test cross as the seed progenies to be evaluated result from inter-crossing between the clones that are under test (i.e. each clone under evaluation is used as female and randomly mated to all, or a good range, of other clonal selections). A polycross, like the test cross, is used to determine the general combining ability of the different potential clonal parents. The seed so produced from a polycross is then tested in randomized field trials. It is essential that the trials are randomized and that the level of replication is high enough to allow the possibility of hybrid seed being from as many other clones as possible.

Seeds from test crossing and polycrossing are grown in progeny evaluation trials to evaluate the genotypic worth or determine the general combining ability of each of the clones. Progeny evaluation trials are very similar to any other plot evaluation trial and are best grown in more than a single environment. Progeny evaluations are also often repeated over a number of seasons to obtain more representative evaluations of the likely performance over different years.

Depending on the results from the progeny evaluation trials, clones that show greatest general combining ability will be used as parents for the synthetic cultivar. These clonal parents are mated in a number of combinations to produce experimental synthetics. The number of clones that are selected and the number of parents that are to constitute the synthetic will determine how many possible combinations are possible. If there were only four clones selected, then there would be a total of 11 different synthetic cultivars (that is, the six possible 2-clone combinations, the four 3-clone combinations and the one 4-clone combination). With 6, 8, 10 and 12 parents, the number of possible synthetics would be 57, 247, 1,013 and 4,083, respectively. Therefore it is useful to try to predict the performance of synthetic cultivars without actually producing seed. One formula used is:

where is the mean performance of all possible single crosses among parents, and is the mean performance of the parents. It should, however, be noted that there is an assumption of an absence of epistasis (interaction between alleles at different loci) in order to obtain a good estimate of synthetic performance, and so predictions are therefore often far removed from what is actually observed. Such predictive methods should therefore be used with caution.

4.5.1 Seed production of a synthetic cultivar

The parents used to produce a synthetic cultivar can be intercrossed by hand-pollination to produce the first-generation synthetic (Syn.1). The aim is to cross every parent in the synthetic with all others (i.e. a half diallel cross). This can be difficult with some synthetics (e.g. alfalfa) where the number of parents included is often around 40.

It is therefore more common in situations where many parents are used in a synthetic cultivar to produce Syn.1 seed using a polycross procedure. The selected parents are grown in close proximity in randomized block designs with high replication. The crossing block is obviously grown in isolation for any other source of the crop to avoid cross-contamination. In cases where insect pollinators are necessary to achieve cross-pollination, attempts are made to ensure that these vectors are available in abundance. For example, alfalfa breeders introduce honey bees or leaf-cutter bees to pollinate synthetic lines.

Seed from Syn.1 is open-pollinated to produce Syn.2, which is subsequently open-pollinated to give the Syn.3 population, and so on. The classes of synthetic seed are categorized as breeders' seed, foundation seed and certified seed. In this case breeders' seed would be Syn.1, foundation seed Syn.2, and the earliest certified seed Syn.3.

In summary, the characteristics of a synthetic cultivar are:

• Synthetic cultivar species need to have potential to have parental lines, which reproduce from source material (either clonally or as an inbred line), and hence populations with the specific genetic makeup of the synthetic cultivar can be reproducibly produced from these base parents.
• To develop synthetic cultivars, the contributing parental material is tested for combining ability and/or progeny evaluated.
• Pollination of synthetic cultivar species cannot be controlled, and there has to be some natural method of random mating between parents (e.g. a suitable out-crossing mechanism such as a self-incompatibility system or other mechanism to promote cross pollination combined with an appropriate pollinator or wind pollination).
• The source parental material must be maintained for further use.
• Open-pollinated populations have limited life and are then reconstituted from the base parental lines on a cyclic basis.

4.6 Developing hybrid cultivars

Crops that are commonly produced and sold as hybrids include maize, Brussels sprouts, kale, onions, canola, sorghum, sunflower and tomato.

Although attempts have been made to develop hybrid cultivars from almost all annual crop species, it is unfair and incomplete to consider the evolution of hybrids without a brief history of the developments in hybrid corn. At the beginning of the 20th century it became apparent that genetic advances and yield increases achieved by corn breeders were markedly lower than those realized by other small-grain cereal breeders developing wheat and barley cultivars. Indeed, the pedigree selection schemes used by corn breeders, although considered suitable at that time, were essentially not effective. The knowledge that hybrid progeny produced by inter-mating two inbred lines often showed hybrid vigour or heterosis (i.e. trangressive segregation whereby the hybrid progeny produces higher yield than the better parent) suggested that hybrid cultivars could be exploited by corn breeders on an agricultural scale by manually detasseling female parents to produce a population that was entirely composed of hybrid seed that could be sold for commercial production.

The first suggestion of using controlled crosses was made by W.J. Beal in the late 19th century, based specifically on the concepts of Darwin on inbreeding and outbreeding. These ideas were then refined by G.H. Shull in 1909, on the basis of genetic studies, who put forward the idea of a hybrid cross being produced by first developing a series of inbred, or near-inbred, breeding lines and inter-mating these as crosses (single cross hybrid) and using the hybrid seed for production. These hybrid progeny were indeed high-yielding and showed a high degree of crop uniformity. His basic concept, however, was not adopted at that time because even the most productive inbred lines had very poor seed yields (most likely due to inbreeding depression) and consequently, hybrid seed production was very expensive.

In the interim the traditional open-pollinated corn cultivars were quickly superseded by double cross hybrids suggested by D.F. Jones in 1918. Double cross hybrids were not as high yielding or as uniform compared with the single cross hybrids proposed by Shull. However, hybrid seed production was less expensive than single cross hybrids, and as a result double cross hybrids completely dominated US corn production by the 1940s.

Initially all commercial hybrid corn seed was produced by detasseling female plants and growing them in close proximity to non-detasseled males, and harvesting seed only from the female plants. When this method of hybrid corn seed production was most prominent it was estimated that more than 125,000 people were employed in detasseling operations in the US in any growing season. Increased labour costs, combined with developments in using cytoplasmic male sterility (CMS) production systems in the 1960s, rendered detasseling of females in hybrid corn production obsolete for some time, but is now once again the method of choice, although some other biotechnology approaches to this are now being considered.

From the onset of hybrid corn breeding it was realized that there was a limit to production based on the inbred lines available, and a big effort was put into breeding superior inbred lines to use in hybrid combinations. Introduction of efficient and effective CMS hybrid production systems combined with the development of more productive inbred parents led to the return of single cross hybrids, which now dominate corn production in the US and many other developed countries. The relatively high cost of hybrid corn and other hybrid crop seeds does, however, limit the use of hybrid cultivars in many developing countries. Nevertheless, in several South American countries like Argentina, Chile, Paraguay, Uruguay and Brazil, nearly all the maize seed used by farmers is now hybrid because the superior yield they provide largely offsets their higher price.

The rapid increase in popularity and success achieved in hybrid maize could not have occurred without two very important factors. The first is that many countries, including until recently the US, do (or did) not have any Plant Variety Rights legislation or other means that breeders could use to protect proprietary ownership of the cultivars they bred. There was, therefore, little incentive for private companies to spend time, resources and effort in developing clonal, open-pollinating or inbred cultivars as individual farmers could increase seed stocks themselves, or they could be increased and sold by other seed companies. Hybrid varieties offered the potential for seed/breeding companies to have an in-built economic protection. The developing companies kept all the stocks of the parental lines and only sold hybrid seed to farmers. These hybrids, although uniform at the stage, would segregate if seed were retained and replanted (i.e. they would be progenies). Secondly, the introduction of hybrid maize occurred simultaneously with the transition from traditional to intensive technology-based agricultural systems. The new hybrids were indeed higher yielding, but were also more adapted to the increased plant populations, rising soil fertility levels and overall improved crop management of the times.

There are hardly any agricultural crops where hybrid production has not at least been considered, although hybrids are used in still relatively few crop species. The reasons behind this are firstly that not all crops show the same degree of heterosis found in maize, and secondly that it is not feasible in many crop species to find a commercial seed production system that is economically viable in producing commercial hybrid seed. Indeed, if maize had not had separate male and female reproductive organs and hence allowed easy female emasculation through detasseling, hybrid cultivar development might never have been developed, or acceptance would have been delayed by at least 20 years, until cytoplasmic male sterility systems were available.

Hybrid cultivars have been developed, however, in sorghum, onions and other vegetables using a cytoplasmic male sterility (CMS) seed production system; in sugar beet and some Brassica crops (mainly Brussels sprouts, kale and canola) using CMS and self-incompatibility to produce hybrid seed; and in tomato and potato using hand emasculation and pollination.

If hybrid cultivars are to be developed from a crop, then the species must:

• show a high degree of heterosis;
• be capable of being managed so as to produce inexpensive hybrid seed;
• not easily be produced uniformly, and have a high premium for crop uniformity.

There are many differing views regarding the exact contribution of hybrids in agriculture. In hybrid maize there would seem little doubt that there have been tremendous advances made. However, this has been the result of much research time and also large financial investments. In addition, it should be noted that the yielding ability of inbred parents in hybrid breeding programmes have been improved just as dramatically as their hybrid products. Most other hybrid crops (with the exception of sorghum) are also outbreeders. Almost all outbreeding crops show degrees of inbreeding depression and, therefore, its counterpart heterosis. In such cases there are strong arguments, certainly in practical terms, for exploiting heterozygosity to produce productive cultivars. This implies that hybrid cultivars can offer an attractive alternative over open-pollinated cultivars or even synthetic lines, although seed production costs will always be a major consideration. In inbreeding crops, hybrid cultivar production is much more difficult to justify on ‘biological grounds’.

Committed ‘hybridists’, of whom there are many (especially within commercial seed companies), would argue that:

• Yield heterosis is there for the exploitation.
• Hybrid cultivars are economically attractive to breeding organizations and seed companies.
• Technical problems (usually associated with seed production) are simply challenges to be faced.
• The biological arguments are irrelevant.
• The rate of hybrid cultivar adoption is overwhelmingly rapid in countries where farmers can afford them.

Skeptics (of which there seem to be fewer, or who are less outspoken) argue on the basis of experimental data available to date that:

• There is no good evidence for over-dominance, and so it is definitely possible to develop pure-line, inbred cultivars that are as productive as hybrid lines.
• The economic attractions of seed companies should be weighed against high seed costs for the farmers; that technical skills could be put to better (more productive) use.
• The biological reality is all-important.

These latter sceptics usually, however, accept that in the case of outbreeding species, hybrids give a faster means of getting yield increases, while in the longer term inbred lines would match them, but in inbreeding crops this differential in speed is not present.

4.6.1 Heterosis

The performance of a hybrid is a function of the genes it receives from both its parents, but can be judged by its phenotypic performance in terms of the amount of heterosis it expresses. Many breeders (and geneticists) believe that the magnitude of heterosis is directly related to the degree of genetic diversity between the two parents. In other words, it is assumed that the more the parents are genetically different, the greater the heterosis will be. To this end, it is common in most hybrid breeding programmes to maintain two or more distinct germplasm sources (heterotic groups). Breeding and development is carried out within each source and the different genetic sources are only combined in the actual production of new hybrid cultivars or while testing experimental combinations. For example, maize breeders in the US observed significant heterosis by crossing Iowa Stiff Stalk breeding lines with Lancaster germplasm. Since this discovery, these two different heterotic groups have not been intercrossed to develop new parental lines, but rather have been kept genetically separated for parental development, so crossing and selection has been imposed on each heterotic group separately.

It has proved difficult to clearly and convincingly define the underlying causes of heterosis in crop plants. There are very few instances where heterozygous advantage per se has been shown to result from over-dominance. The counterpart to heterosis, inbreeding depression, is generally attributed to the fixation of unfavourable recessive alleles, and so it is argued that heterosis should simply reflect the converse effect. Therefore unfavourable recessive alleles in one line would be masked, in the cross between them, by dominant alleles from the other. If this is all there is to it, then heterosis should be fixable in true breeding lines by the selection of lines with only the favourable alleles. In general it has been found that this simple rationalization does not explain all the observed effects. Thus, the question is whether this breakdown in the explanation is related to a statistical problem of the behaviour of a large number of dominant/recessive alleles, each with small effect; whether the failure to detect over-dominance is simply a technical failure rather than a lack of biological reality; or whether a more complex explanation needs to be invoked. Dominance can be regarded as the interaction between alleles at the same locus, their interaction giving rise to only one of their products being observed (dominance is expressed) or a mixture of the products of the two (equal mixing giving no dominance, and inequality of mixing giving different levels of incomplete dominance). But another well-established type of interaction of alleles can occur, that between alleles at different loci (called non-allelic interaction or epistasis). In addition we cannot simply ignore the fact that linkage between loci is a recognized physical reality of the genetic system that we now regard as being the basis of inheritance. It has been shown that the combination of these two well-established genetic phenomena can produce effects that are capable of mimicking over-dominance.

To examine the effect of having many loci showing dominance and recessivity, determining the expression of a character differing between the parents of a single cross, let us examine the case of two genetically contrasting parental lines that differ by alleles at only five loci. Consider the two following cases, where we assume that: capital letters represent ‘increasing’ alleles, lower case ‘decreasing’ ones; each locus contributes in additive fashion to the expression of the character (i.e. the phenotype, let us consider yield); each increasing allele adds the same amount to the yield (2 units) while the decreasing allele adds nothing to yield;

---

Case 1

and then:

	Parent 1		Parent 2
Genotype Phenotype (yield)	AABBCCDDEE		aabbccddee
Genotype Phenotype (yield)		AaBbCcDdEe

---

and that dominance is complete and for increasing expression.

The level of heterosis can be considered in two forms: one is the minus the mid-parent, called mid-parent heterosis, while the second is minus the best parent, called best parent heterosis. The mid-parent heterosis need not detain us here, as it is of no direct interest. If best parent heterosis is positive (i.e. the exceeds the performance of the best parent), then generally heterosis has been ascribed to the presence of over-dominance, but this need not be the case.

In Case 1 the best parent has a phenotype of 10 yield units; and the has a phenotype of 10 yield units; that is, they have identical yields and no heterosis is detected.

In Case 2 (with exactly the same alleles and effects but with different starting arrangement of alleles between the parents), the best parent has a phenotype of 6 yield units; and the has a phenotype of 10 yield units; that is, heterosis would be 4 units and the is, in fact, 40% more productive than the better parent.

It is, however, clearly wrong to consider this to be over-dominance, since none of the individual loci show such an effect. It would certainly be possible (in fact, Parent 1 in Case 1 shows this) to fix this level of performance in homozygous, true breeding lines with no heterozygosity (i.e. AABBCCDDEE).

Now let us consider the statistical probability underlying these combinations. With five loci, assuming no linkage, no effects of selection and a random assortment of gametes in the , after a

---

Case 2

and then:

	Parent 1		Parent 2
Genotype Phenotype (yield)	AAbbCCddEE		aaBBccDDee
Genotype Phenotype		AaBbCcDdEe

---

number of rounds of selfing, the probability (which is equivalent to the frequency) of having a genotype that combines the five dominant alleles as homozygotes would be (or 0.03125, just over 3%). If, however, you wish to have some assurance that a breeding population contains at least one of these genotype recombinants, amongst all the possible inbred lines that can be produced, the number of lines needed to be grown and screened is given by the equation:

where is the number of inbred lines that would need to be screened, is the desired probability that at least one line with the desired genotype will exist in the population, and is the frequency of the genotype of interest. In the example with five loci, it would be necessary to screen at least 145 inbred lines from the progeny derived from the cross between Parent 1 and Parent 2 (in both cases) to be 99% certain that at least one example of the genotype required would exist.

Of course, a quantitatively inherited trait like yield is not controlled by five loci but more likely 50, 500 or 5,000. To give some idea of the number of genotypes needed to obtain a specific combination of alleles, when the number of loci increases, see Table 4.1. The number of plants that need to be screened to be 90%, 95%, 99%, 99.5% and 99.9% sure that at least one genotype of the desired combination exists, is shown for cases with 5, 6, 7, 8, 9, 10 and 20 loci. With 20 loci, a modest number compared with the possible real situation, a breeder would need to evaluate almost 5 million inbred lines to be sure that the one he wants is present. So hybrids do offer a better probability of success in this instance, but not because they show over-dominance at their loci!

Table 4.1 Number of recombinant inbred lines that would require evaluation for the breeder to be 90%, 95%, 99%, 99.5% and 99.9% sure that a homozygous lines will be a specific combination of alleles at each of 5, 6, 7, 8, 9, 10 and 20 loci.

	Probability of obtaining at least one individual of the desired genotype
	0.900	0.950	0.990	0.995	0.999
5 loci	76	94	145	167	218
6 loci	146	190	292	336	439
7 loci	294	382	587	676	881
8 loci	588	765	1,177	1,354	1,765
9 loci	1,178	1,532	2,356	2,710	3,533
10 loci	2,357	3,066	4,713	5,423	7,070
20 loci	2,414,434	3,141,251	4,828,869	5,555,687	7,243,317

A second consideration with hybrid lines that has been postulated is that the heterozygosity of the cultivar makes it more stable over a range of different environments. This may be true, but there is no direct evidence for such a basic biological effect and it does not explain the extremely high genotype by environment interactions found to be exhibited by hybrid maize cultivars.

4.6.2 Types of hybrid

There are a number of different types of hybrid, apart from the single cross types concentrated on above. The different types of hybrid differ in the number of parents that are used in hybrid seed production. Consider four inbred parents (A, B, C and D) – types of hybrid that could be produced would include:

• Single cross hybrids (e.g. etc.)
• Three-way hybrids (e.g. etc.)
• Double cross hybrids (e.g. etc.)

Single cross hybrids are genetically uniform, whereas three-way or double cross hybrids are genetically heterogeneous (a three-way cross is less heterogeneous that a double cross). In general single cross hybrids have the highest level of heterozygosity and are more productive than the other two, but on the other hand are most expensive in terms of hybrid seed production.

4.6.3 Breeding system for hybrid cultivars

The three major steps in producing hybrids are therefore:

• development of inbred lines to be used as parents;
• test cross these lines to identify those that combine well;
• exploit the best single crosses as hybrid cultivars.

The system used to develop hybrid cultivars is illustrated in Figure 4.7. The scheme involves six stages:

• Produce two, or more, segregating populations.
• Develop inbred lines (parents) from each of the two populations.
• Evaluate performance of inbred lines phenotypically.
• Evaluate general combining ability of selected inbred lines.
• Evaluate hybrid cross combinations.
• Increase inbred parental lines, and commercialize hybrid cultivars.

Figure 4.7 Outline of a hybrid breeding scheme.

The procedure used to develop inbred parent lines in hybrid cultivar development is similar to that used to breed pure-line cultivars, and the advantages and disadvantages of various approaches are the same. Breeders have used bulk methods, pedigree methods, bulk/pedigree methods, single seed descent and out-of-season extra generations (off-station sites) to achieve homozygosity. Regardless of the breeding approach of choice, these are independently imposed on the different heterotic groups from which inbreds are extracted. One of the most important objectives is to maintain high plant vigour and to ensure that the inbred lines have as high seed productivity as possible. This is not always easy, particularly in species where there is a high frequency of deleterious recessive alleles present in the segregating populations. Breeders must decide the level of homozygosity that is required. On one hand, the more homozygous (the extreme, of course, being 100% homozygosity) the inbred lines are, then the more uniform will be the resulting hybrid. The more heterozygous inbred lines may, however, be more productive as parents and hence help to reduce the cost of hybrid seed production.

Combining ability (or more relevantly, general combining ability, GCA) is evaluated with the aim of identifying parental lines which will produce productive progeny in a wide range of hybrid cross. Generally, it is not possible to cross all possible parental lines in pairwise combinations, as the number of crosses to be made and evaluated increases exponentially with the increased number of parents. It is therefore more usual to cross each parent under evaluation to a common test parent or tester. The tester used is common to a set of evaluations, and therefore, general combining ability is determined by comparing the performance of each progeny, assuming that the only difference between the different progenies can be attributed to the different inbred parents. Testers are usually highly developed inbred lines that have proved successful in hybrid combinations in the past. A far better prediction of general combining ability would be achieved if more than one tester were used. This, however, is not common practice, and breeders have tended to prefer to test more inbreds rather than to increase the number of test parents used. As the hybrid programme progresses, it is common, however, to replace older testers with newer ones.

Evaluation of specific combining ability (or actual individual hybrid combination performance) is carried out when the number of parents is reduced to a reasonable level. The number of possible cross combinations differs with the number of parents to be tested. The number of combinations is calculated from:

where is the number of parents to be evaluated. For example, if 20 parents are to be tested then there would be: 20 crosses to a single tester; 190 pairwise crosses, 3,420 three-way crosses and 14,535 double cross combinations possible. It is therefore common to predict the performance of three-way and double crosses from single cross performance rather than actually test them. The three-way cross performance can be predicted from the equation:

where is the performance of the progeny from the cross between and . It is noted that the actual single cross in the hybrid predicted is not used in the prediction.

To predict the performance of a double cross the following equation is used:

where is the performance of the progeny from the cross between and , and so on. Note again that the two single crosses used in the double cross do not appear in the prediction equation.

The assumptions underlying this will not be discussed here, but we simply note that this is what is carried out quite often in practical breeding.

4.6.4 Backcrossing in hybrid cultivar development

Backcrossing has featured quite highly in hybrid breeding schemes. Backcrossing is used in hybrid development for two purposes:

• To introduce a single gene into an already desirable inbred parent.
• To produce near isogenic lines, which can reduce the cost of seed production, by convergence, called convergent improvement. This is used to make slight improvements to specific hybrid cross combinations and involves recurrent backcrossing between a single cross hybrid and both of its two parents ( and ). The result will be to produce, from (where is a near-isogenic line of ). Similarly, near-isogenic lines are developed for . These can be used in a modified single cross (; see also A' maintainer lines in cytoplasmic male sterile systems, below), or a double modified single cross . The aim of this is to increase the efficiency and reduce the cost of seed production.

4.6.5 Hybrid seed production and cultivar release

The inbred lines used as parents are increased in exactly the same way as pure-line cultivars, and hence no further description is needed here.

The first and highest priority of hybrid seed production is to complete the task as cheaply as possible with the maximum proportion of hybrid offspring. There are four basic means that have been used to produce commercial amounts of hybrid seed:

1. Mechanical production In hermaphroditic plants, females are emasculated and pollination is achieved either by hand or naturally. In diclinous plants simple physical protection (such as bagging) and hand-pollination are used. The problems with mechanical hybrid production are related to cost, labour inputs and the number of seeds produced by each pollination operation. Examples include tomatoes and potatoes, along with detasseling (as described earlier) in maize.
2. Nuclear male sterility Nuclear male sterility (NMS) is found in many crop species, and sterility is usually determined by the expression of recessive alleles (ms ms) while fertility is present with the dominant homozygote (Ms Ms) and the heterozygotes (Ms ms). Clearly it is not possible to maintain a pure male sterile line by simple sexual reproduction. However, if seeds are collected from male sterile plants they will be heterozygous for the sterility locus and on selfing will segregate 75% fertile (Ms ms or Ms Ms) and 25% sterile (ms ms). The major problems with NMS are related to introducing the necessary genes into commercially desirable genotypes, and the expression being environmentally dependent. Success has been achieved in castor and attempts have been made with cotton, carrot, tomato, sunflower and barley.
3. Self-incompatibility To produce hybrids one simply interplants two self-incompatible, but cross compatible, inbred lines and harvests all the seed, all of which will be hybrid. Problems have been related to overcoming the self-incompatibility of parental lines in order to maintain them, which can be costly (e.g. bud pollination by hand is often necessary to overcome incompatibility in some Brassica species). Incompatibility systems are also often influenced by environmental conditions. Examples of hybrid crops include several vegetables, canola and sunflower.
4. Cytoplasmic male sterility Cytoplasmic male sterility (CMS) is of wide natural occurrence and can sometimes be induced by backcrossing a desired, adapted genotype (i.e. its nucleus) into a foreign cytoplasm such that the nuclear–cytoplasmic interaction results in male sterility. A practical hybrid system also requires, of course, the restoration of male fertility, usually achieved with ‘restorer genes’. These, fortunately, are mostly dominant. If the hybrid product need not produce seeds (e.g. onions or beets) the restorer gene is then not needed. The combination of effective sterility and reliable restorer genes are not often available. Another problem is that the cytoplasm itself (which is always a component of the resulting hybrid) often has deleterious effects on plant vigour. The lack of pollen in the female plants can also have an effect on hybrid production. For example, if the pollinating vectors are insects they may well prefer to visit only the fertile males. Examples of success are onions, carrot, sugar beet, maize and sorghum.

To produce hybrid cultivars using cytoplasmic male sterility requires three types of genotype: male fertile lines (called B lines) with no cytoplasmic male sterility genes (Figure 4.8), which are homozygous for a dominant restorer gene (i.e. normal cytoplasm, Rf Rf); cytoplasmic sterile female lines with male sterile cytoplasm (cms) with no restorer genes, rf rf (called A lines); and ‘male-fertile’ female lines (called A' lines or A' maintainer lines) with normal cytoplasm and no restorer genes, rf rf. A' lines are usually isogenic lines of the cytoplasmic male sterile A-female parents, produced by recurrent backcrossing, and are used to maintain and increase seed of the A cytoplasmic male sterile female parent.

Figure 4.8 Hybrid seed production system using cytoplasmic male sterility requires three types of genotype: male fertile lines (called B lines) with no cytoplasmic male sterility genes and which are homozygous for a dominant restorer gene; cytoplasmic sterile female lines with male sterile cytoplasm but with no restorer genes; and ‘male-fertile’ female lines (called A' lines or A' maintainer lines) with normal cytoplasm and no restorer genes.

In addition, several companies have used biotechnology to develop systems that enable inbred lines to become male sterile through the expression of genes specifically in male reproductive tissues. These will become commercially available in the years to come.

4.7 Development of clonal cultivars

Crops that are generally produced as clonal cultivars include bananas, cassava, citrus, potatoes, rubber trees, soft fruit (raspberry, blackberry and strawberry), sugarcane, sweet potatoes, and top fruit (apples, pears, plums, etc.).

Clonal crops are perennial plant species, although a few clonally propagated crops (e.g. potato and sweet potato) are grown as annual crops. Some clonally propagated crops are very long-lived (e.g. rubber, mango and rosaceous top fruits) and can produce crops for many years after being established. Indeed, there are instances whereby fig and palm cultivars have survived over a thousand years, and are still in commercial production. Other clonally propagated crops have a shorter lifespan yet remain in commercial production for several years after being propagated (e.g. sugarcane, bananas, pineapples, strawberries and Rubus spp).

There are many methods of propagation used in clonal crop production. Apples, pears, cherries, various citrus, avocados and grapes are propagated through budding and grafting onto various root stocks. Leafy cuttings are used to propagate pineapple, sweet potato and strawberry. Leafless stem cuttings are used to propagate sugarcane and lateral shoots are used for banana and some palms. There also is, for a number of species, the potential for clonal reproduction via tubers (swollen stems), as is the case for potatoes.

In general, clonal crop species are out-breeders, which are intolerant to inbreeding. Individual clones are highly heterozygous and so it is easy to exploit the presence of any heterosis that is exhibited. Imagine, for example, that corn could be easily reproduced asexually (say through apomixis); then there would be no need to develop hybrid corn cultivars because the highly heterozygous nature of a hybrid line could be ‘genetically fixed’ and exploited through asexual reproduction.

The process of developing a clonal cultivar is, in principle, very simple. Breeders generate segregating progenies of seedlings, select the most productive genotypic combination and simply multiply asexually; this also stabilizes the genetic makeup (i.e. avoids problems relating to genetic segregation arising from meiosis). Despite the apparent simplicity of clonal breeding, it should be noted that while clonal breeders have shared in some outstanding successes, it has rarely been due to such a simple process, as will be noted from the example below.

4.7.1 Outline of a potato breeding scheme

The breeding scheme that was in use at the Scottish Crop Research Institute prior to 1987 is illustrated in Figure 4.9. This breeding scheme is similar to what many potato breeders still use today. It should be noted that the programme used two contrasting growing environments. The seed site (indicated by grey boxes) was located at high altitude and was always planted later and harvested earlier than would be considered normal for a typical ware crop (indicated by black boxes), in order to minimize problems of insect-borne virus disease infection. A ware crop is the crop that is produced for consumption rather than for replanting.

Figure 4.9 Potato breeding scheme used at the Scottish Crop Research Institute prior to 1987.

Each year, between 250–300 cross pollinations were carried out between chosen parents. From each cross combination the aim was to produce around 500 seeds, leading to about 140,000 seedlings being raised in small pots grown in a greenhouse (two greenhouse seasons were needed to accommodate the 140,000 total). At harvest, the soil from each pot was removed and the tubers produced by each seedling were placed back into the now empty pots. At this stage a breeder would visually inspect the small tubers in each pot (each seedling being a unique genotype) and either select or reject the produce of each seedling. One tuber was taken from amongst the tubers produced by the selected seedlings, while all tubers from rejected clones were discarded. The seedling generation in the greenhouse, as in most clonal crops, was the one and only generation that derived directly from true botanical seeds (i.e. from sexual reproduction). All other generations in the programme were derived by vegetative (clonal) reproduction, in other words from tubers.

The following year the single selected tubers (approximately 40,000) were planted in the field at the ‘seed site’ as single plants within progeny blocks. This stage was referred to as the first clonal year. At harvest each plant was hand-dug, and the tubers exposed on the soil surface in a separate group for each individual plant. A breeder would then visually inspect the produce from each plant and decide, on that basis, to reject or select each group of tubers (i.e. each clone). Three tubers were retained from selected plants and planted in the field at the same seed site in the following year (the second clonal year) as a three-plant, unreplicated plot. First and second clonal year evaluations were therefore carried out under seed site conditions to reduce as far as possible the chances of contamination of the plants and hence the tubers, especially by virus diseases.

Second clonal year plots were harvested mechanically and, again, tubers from each plot exposed on the soil surface. A breeder examined the tubers produced and decided to select or reject each clone, again on the basis of visual inspection. Tubers from selected clones were retained and grown in the subsequent year for the first time under ‘ware growing conditions’ in unreplicated trails at the third clonal year stage. Tubers from each selection were also retained at the seed site and used to grow a six-plant plot in the following year. After the second clonal year, however, the seed site was only used to increase clonal tubers and no selection was carried out on the basis of the performance at this site.

At the ware site, measurements of a variety of characters were taken and yield was recorded (Figure 4.10). Thus the third clonal year was the first one where selection was based upon objective measurements, principally yield but also other performance characters and disease reaction. The fourth and fifth clonal generations were repeats of the third year with reduced numbers of entries after each successive round of selection, but with more replicates and larger plots; this was paralleled by larger multiplication plots of 20 plants and 100 plants, respectively, at the seed site.

Figure 4.10 Potato yield assessment trials produce a vast volume of product (tubers) from a relatively small area.

In the sixth, seventh and eighth clonal generations, surviving clones were evaluated at a number of different locations (‘regional trials’) in the UK. After each round of trials, the most desirable clones were advanced (i.e. re-trialled) and less attractive clones discarded.

Clones that were selected in each of the three year's regional trials were entered into the UK National List Trials (a statutory government-organized national testing scheme). Depending on performance in these trials a decision was made regarding cultivar release and initial foundation seed lots were initiated. If all went well, farmers could be growing newly developed cultivars within 17 years of the initial cross being made!

4.7.2 Time to develop clonal cultivars

Despite the lengthy time period between crossing and farmers growing a new potato cultivar, this is a short time period in comparison with some of the other asexually propagated crops. In potato the long selection process is related to the difficulty in evaluating a crop where the phenotype is greatly affected by the environment (both where the seed and ware crops were grown). In the case of potato, some of the length of the process is related to a slow multiplication rate, around 10:1 per generation. In addition, seed tubers are bulky and require large amounts of storage space. Planting material for one acre of potatoes will require approximately 2,000 lbs of seed tubers. To some extent this lengthy period can be reduced by in vitro multiplication of selected genotypes, but this does not reduce the need for adequate testing of genotypes over a reasonable number of seasons and sites.

With many other clonal species the process from crossing to cultivar release can be very lengthy. In apple breeding, for example, it is often said that if a breeder is successful with the very first parent cross combination, then it is still unlikely that the cultivar will be released (from that cross) by the time the breeder retires! In this case there is the obvious difficulty in the time taken from planting an apple seed to the time that the first fruit can be evaluated. With several other clonal crops, even the time to develop a new apple cultivar can be considered a relatively short time period in terms of the life of a cultivar. For example, the most common date palm cultivar in commercial production is a clonal line derived from the Middle-east called ‘Siguel’. To our knowledge, it is so old that no one knows the derivation of this line, but it still predominates as the leading cultivar around the world.

Clonal crop species have shown a high frequency of exploited natural mutations compared with other crop types, probably simply as a result of their clonal propagation making any ‘variant’ obvious and easily multiplied. As a result many clonal cultivars are the result of natural mutations rather than arising from selection following a specific hybridization between parental lines. For example, the potato cultivar ‘Russet Burbank’ is a mutation from ‘Burbank’, and similarly ‘Red Pontiac’ is a natural mutation of ‘Pontiac’, and many apple cultivars are simply fruit colour mutants.

4.7.3 Sexual reproduction in clonal crops

In crops that are reproduced from true botanical seed, there is definite selection for reproductive normality and high productivity of sexual reproduction. In clonal crops this has not always been the case, and the result can hinder the ability of breeders to generate variation by sexual reproduction. There are two main types of clonally reproduced crops (excluding apomicts):

• those that produce a vegetative product;
• those producing a reproductive product (i.e. fruit).

Those that produce a vegetative product have almost all been selected to have reduced sexual reproductive capacity and can exhibit problems in relation to sexual crosses. This is probably the result of conscious or subconscious selection for plants that do not ‘waste energy’ on aspects of sexual reproduction and will therefore put more energy into the vegetative parts. Extremes are found in yams and sweet potatoes in which many cultivars never flower and in many cases they cannot be stimulated to reproduce sexually. Modern potato cultivars have far less flowering than their wild relatives, and of those that do flower, many have very poor pollen viability or are pollen sterile. In the case of potato this has been the result of conscious human selection, based on the argument noted above and where sexual seed development can also cause a problem of generating ‘volunteers’ in subsequent crops.

In clonal crops where the reproductive product is used, there is of course no question about reducing flowering, but nonetheless, reproductive peculiarities and sterility problems are still very common. In general, selection has favoured the vegetative part of fruit development at the expense of seed production. In the extreme, bananas are vegetatively parthenocarpic (i.e. formation of fruit without seeds). Wild bananas are diploid and reproduce normally, producing fruit with large seeds. Commercially grown bananas are triploid and hence sterile, so many banana cultivars cannot be used as parents in a breeding scheme. Pineapples are also parthenocarpic but self-incompatible, so that clonal plantings give seedless fruits, even though fruit would be seedy if pollinated by another genotype. In addition, mangoes and some citrus sometimes produce polyembryonic plants (where multiple embryos are formed from a zygote by its fission at an early development stage, and hence in effect, results in clonal pseudo-seedlings identical to the mother genotype), which can be a great nuisance to breeders when they are selecting amongst segregating populations of sexually generated progeny.

In conclusion, reproductive derangement in clonal species can result in potential sexual crosses between particular parent combinations not being possible, or that individual parental lines cannot be used for sexual reproduction. This limits the options open to breeders of clonally reproduced crops.

4.7.4 Maintaining disease-free parental lines and breeding selections

Several decades ago it was thought that clonal cultivars were subject to clonal degeneration, seen as a reduction in productivity with time. It is now known that this degeneration is primarily the result of clonal stocks becoming infected with bacteria or viruses. For example, both potatoes and strawberries can be affected in this detrimental way by infection by bacterial or viral diseases. As an additional example, a bacterial disease causes stunting disease in sugarcane ratoons, which is not, as was postulated, caused by genetic ‘drift’ bringing degeneration. It should, however, be noted that the cause of reduction in performance is often not solely due to disease buildup, and some degeneration of clonal lines can be the result of accumulation of deleterious natural somatic mutations.

Nevertheless, a primary problem in breeding clonal crops is to keep parental lines and breeding stocks free from viral and bacterial diseases that are transmitted through vegetative propagation. Often it is necessary to maintain disease-free breeding stocks that are separated from those being tested for adaptation potential, as was the case with the potato breeding scheme described above.

4.7.5 Seed increase of clonal cultivars

One of the primary concerns associated with the increase of new clonal cultivars is to prevent the stocks becoming infected by viral or bacterial diseases during commercial increase. Biotechnology approaches such as in vitro culture can often help in providing rapid increases of clonal stocks, which can be readily maintained true to type. They also enable the elimination, or at least a very significant reduction, of the pathogen load that builds up through continuous clonal propagation in the field. Such methods, however, cannot always be easily applied to many (particularly woody plant types) species, where rates of multiplication in vitro are often low and re-establishment under normal field conditions is not always possible. In vitro multiplication and applications for disease-free clonal material are discussed later.

4.8 Developing apomictic cultivars

It is possible to develop clonal cultivars through obligate apomixis (e.g. as with buffel grass, Cenchrus ciliaris). In species that are grown from obligate apomictic seeds, there tends to be a positive relationship between the productivity of a clone and its level of heterozygosity. It is therefore desirable to develop sources of genetic variability to maximize productivity within the population and this is a continual challenge to breeders of apomictic crops. Essentially all the seed produced by an obligate apomict are produced as a result of asexual reproduction. By far the most common means of producing genetic variation and/or increasing heterozygosity, however, is by sexual reproduction or hybridization between two chosen lines or populations. In developing apomictic cultivars it is important that some seed can be sexually produced so that this type of variation can be exploited.

There are basically two methods of producing apomictic cultivars.

• Selfing a sexual clone to produce a segregating (due to the clone being highly heterozygous). The progeny can be screened for ‘sexual’ and ‘apomicts’. If apomicts are obligate they have the potential of being a clonal cultivar.
• Crossing a sexual clone to an apomict clone to produce an ‘’ and selecting for sexual or apomicts reproducing clones within the segregating population. Again, obligate apomicts have the potential of being new cultivars.

The method is illustrated in Figure 4.11. Of course, a breeding scheme can use a combination of both methods of selfing and intercrossing to develop a breeding system. In many cases lines have been developed by first selfing clones to achieve some degree of inbreeding, which is aimed at reducing the frequency of deleterious alleles, and then intercrossing the partial inbred lines to develop the apomict cultivar. The scheme shown in Figure 4.11 also could begin by selfing a semi-apomictic selection rather than a cross between an apomict and a sexual parent.

Figure 4.11 Breeding scheme for developing near obligate apomict cultivars.

Once the apomicts have been identified, the most adapted lines are selected through repeated rounds of evaluation in different environments and years.

4.9 Summary

In summary, different crop species are more or less suited for developing different cultivar types. Individual plant breeders have developed specific breeding schemes that best suit their situation according to the type of cultivar being developed, the crop species and the resources (financial and others) that are available.

Irrespective of which exact scheme is used, breeding programmes also differ in the number of individual phenotypes that are evaluated at each stage and in the characters that are used for selection at each of the breeding stages. To address the issue of numbers in the breeding programme, it is necessary to consider the mode of inheritance of the factors of interest, and to have an understanding of the genetics underlying their inheritance.

Think questions

1. Disease resistance to powdery mildew has been identified in a primitive, uncultivated species related to cultivated barley (Hordeum vulgare). Research has shown that the resistance is controlled by a single dominant allele . There is a resistance screen test available which shows 100% reliability. Outline a breeding scheme that could be used to introgress this character into an existing cultivar (‘Golden Promise’) which is susceptible to powdery mildew . The aim is for the resulting resistant lines to comprise at least 96% of the Golden Promise genotype.

   Consider now that the disease resistance is controlled by a single recessive allele – how would this affect the breeding method you have described above?

2. Crop cultivars can be divided into pure-line cultivars, open-pollinated populations, hybrid cultivars or clonal cultivars. Briefly, describe the major problems that can be encountered or the attributes of the crop types that can be utilized, when breeding: a pure-line cultivar; an open-pollinated population; a hybrid cultivar; and a clonal cultivar.
3. You have been appointed as alfalfa (Medicago sativa L.) breeder to develop crops for the Alfred Alfoncia Seed Company Ltd. Outline a suitable breeding scheme you will use to develop superior synthetic cultivars for your new employer. Show each stage of the scheme on a year-by-year basis.

   Now, after several rounds of clonal selection, you have identified four potential highly productive parent lines for developing a new synthetic (lines are coded as A, B, C and D). Hybrid seed was produced from all two-parent cross combinations possible and these were evaluated in yield trials along with the parent lines. From the yield (kg) results (below), indicate the three parent synthetic most likely to be highest yielding and state the expected yield.

   
4. Outline three characteristics of a crop species that would merit consideration before developing hybrid cultivars from that species.
5. Outline the advantages and disadvantages of a bulk breeding scheme and a pedigree breeding scheme to develop pure-line cultivars

   Many breeding schemes for pure-lines are a combination of bulk and pedigree schemes. Design a suitable bulk/pedigree breeding scheme (it is not necessary to make notes on when specific characters are selected) that could be used to develop superior pure-line cultivars. Outline any advantages or disadvantages of your breeding scheme.

6. A hybrid breeding programme has identified six superior inbred lines (A, B, C, D, E and F) which were intercrossed in all combinations and the progenies evaluated for productivity in a field yield trial. The results from each single cross in the trial were:

   Which parent combination will provide the most productive single cross hybrid and what would the expected yield be? Which parent combination will provide the most productive three-way cross hybrid and what would the expected yield be? Which parent combination will provide the most productive double cross hybrid and what would the expected yield be?

   Explain why the most productive double cross combination (from the prediction equation) may not result in the most commercially suited hybrid cultivar.

7. In order to justify development of hybrid cultivars it is necessary to have some method of producing inexpensive hybrid seed. Outline three systems that have been used to produce hybrid seed and indicate the advantages or disadvantages of each system.
8. ‘Russet Burbank’ has been the predominant potato cultivar in the Pacific Northwest for many years. You have been asked to design a breeding programme that will produce cultivars that can replace some of the Russet Burbank acreage. Outline the breeding design you would use, indicating: (a) five breeding objectives; (b) the breeding scheme (use diagrams if necessary); and (c) indicating the number of clones evaluated and selected at each stage.
9. In order to produce a hybrid cultivar using cytoplasmic male sterility you require three types of genotypes. Describe the differences between the three types required, and describe how commercial production will be conducted.
10. You are to embark on a backcrossing study designed towards convergence breeding. You have at your disposal two genotypes coded C and M. Genotype C is a cytoplasmic male sterile line, but one that does not have very good hybrid combining ability; genotype M has great hybrid combining ability and is male fertile (without a male fertile restorer gene). Design a scheme (listing each operation necessary) that will result in a genotype which is 95% the nuclear genotype of the M line but is cytoplasmically male sterile.
11. You are walking along a deserted beach in southern California (are there any of these left?) and you inadvertently kick a brass oil lamp. On picking up the lamp you gently rub the side and – poof! – a 20 foot tall genie appears! ‘Oh master! ’ says the genie, ‘I can grant you two wishes’ (well, times are hard and the old three wishes rule does not apply any more). You think for a few minutes, then say ‘I would like you to conjure me up the most wonderful plant that exists within the Universe, so that I can grow it on my farm in Idaho and make millions of dollars in profits’. Poof! This plant appears before you – a plant species that has never been seen on Earth before.

    As a plant breeder, list five questions you would ask the genie about the biology of this plant that would help you to design a cultivar development programme to increase the value of it as a crop.

    Having sorted that out, you begin to think of your other wish. ‘I would like you to tell me the formula for a chemical apomict’ (a chemical that, when applied to a crop, will result in 100% apomictic seed from the crop). ‘Well’ says the genie. ‘I can do this, but you must specify which crop species the chemical will work on’ (it can only work on one chosen species). What crop species would you choose to have apomictic seeds, and why?