Tree improvement takes advantage of the natural genetic variation which exists within a species. A number of improvement techniques are available which exploit this variation at different levels.

Seed origin or provenance testing

Most tree species used in British forestry show large differences in the rate of growth or in form throughout their natural distribution. An essential prerequisite of more sophisticated methods of tree improvement is to investigate this basic variation within a species by comparing range-wide collections of material in experiments planted at a number of relevant test sites. This is the process of seed origin or provenance testing which was started by the Forestry Commission in the 1920s. Since then over 400 experiments have been established consisting of all the major and a number of minor coniferous species and several broadleaved species. However, it is interesting to note that this work, fundamental to effective tree breeding, remained with the Silviculture Branches until the mid-1980s to allow the geneticists to concentrate on breeding activities.

Tree breeding

Within the broad differences attributable to origin, considerable variation exists among individual trees and the more detailed techniques of tree breeding exploit variation at this level. Breeding work seeks to establish whether the observed superiority of an individual is passed on through seed to its progeny in the next generation. The most basic form of improvement is based on this assumption by using superior stands of trees as sources of seed. Stands of an appropriate origin are selected and managed for seed collection from the better individuals within the stand. Use of seed stands quickly results in well-adapted seed but is not likely to be associated with high levels of improvement.

Using plus trees

The tree breeder can achieve higher gains, however, by concentrating on the selection of superior individuals which are referred to as plus trees. The degree of superiority which must be shown by a plus tree with respect to the mean of the population from which it is selected is a reflection of the selection intensity used, a higher threshold reflecting a higher intensity. The collection of plus-trees identified at the beginning of a breeding programme is referred to as the base population.

Establishing whether the observed superiority of a plus-tree is inherited by its progeny is much longer-term, but it can bring considerably greater rewards. When an individual seed collection is made from a plus tree, the seeds collected or the progeny subsequently grown from it are referred to as a family. Because the family members have only the female parent in common (the flowers being wind pollinated with a pollen mixture from surrounding sources) they are half-siblings and this is referred to as a half-sib family. The families of a number of trees are compared against material grown from a standard commercial seedlot in forest stage progeny tests at relevant test sites throughout the country.

These experiments allow the breeder to identify the parent trees whose previously observed superiority is actually inherited by their progeny. The breeder can then use these parents to produce improved seed, and because progeny performance influences parental choice, this is known as backward selection. Plus-trees identified in this way for further use in breeding work are recognised as members of the breeding population. This work of progeny-testing is a very important part of the tree-breeding process involving long-term commitment and making heavy demands on resources. It enables the breeder not only to identify superior parents, but also to study the types and levels of genetic variation present and to determine heritability, the degree to which particular traits (eg height, volume, form) are inherited.

Vegetative propagation

Genetically identical clonal representatives or ramets of one original tree (sometimes called the ortet) may be created by vegetative propagation. Once a parent has been identified as superior in progeny tests it can be propagated in this way, usually by grafting, which maintains some of the mature characteristics of the parent tree. Many grafts can be planted with those of other superior parents in a clonal seed orchard where their genetically superior progeny will be reproduced and can be collected as commercial quantities of seed. The grafts of the different parents are planted in an intimate mixture to promote the maximum amount of intercrossing between parents. Orchards can be constructed to emphasise different selection traits or combinations. Plus-trees used to produce a specific orchard are identified as a production population.

Creating an untested orchard

The procedure described so far can involve a period of at least 20 years between plus tree selection and production of commercial seed from an orchard. This period can be halved by creating an untested orchard established using grafts made at the same time as the parent trees were selected. The parent trees in the orchard would be untested at this stage but the expected improvement to be gained in seed from this orchard would be slightly greater than that obtainable from a registered seed stand because the selection criteria for plus trees will be more intensive and also because both parents are phenotypically superior, whereas the male parents for seed stand progeny are unknown. Information on the performance of the parents in progeny tests could later be used to cull inferior parents from an untested orchard, a process known as rogueing, but the risks of reducing the orchard to too small a number of parents are high if a large number of potential parents in progeny tests emerge as unacceptable. In contrast, much higher gains are to be expected from an orchard based on tested parents.

Seedling seed orchards

For species such as pines, which flower at an early age, seedling seed orchards can be planted at the same time as progeny tests. Seedling progeny of plus trees are planted as the component material of these orchards rather than grafts. They can later be thinned to the best progeny based on evaluation data.

Clone banks

Grafting is also used to propagate material from individual plus trees for systematic archiving. This has the advantage of bringing together a number of ramets from a range of plus trees which are likely to have had a wide geographic distribution. Clonal archives established for this purpose are commonly called clone banks.

Family-mixtures

Some species have shown themselves to be amenable to bulk propagation by cuttings. The quantity of seed required to raise initial stock-plants for this process is comparatively small and can be based on a number of families and reproduced from orchards or clone banks by the tree breeder using artificial pollination techniques. Such work can often be carried out before an orchard is sufficiently mature to produce commercial quantities of seed and during the course of a progeny testing programme. Family-mixtures can therefore be based on the best known parents identified by the most recent progeny test information.

Potential genetic gain

The achievements of tree breeding will ultimately be dependant on three factors:

• The amount of variation present in the population in which selection is being made
• The selection intensity imposed
• The heritability of the selection trait.

These combine to enable the breeder to calculate the potential genetic gain from a breeding programme. Sometimes it will be of interest to verify this by comparing actual commercial products of orchards or family mixtures with unimproved material in genetic gain trials.

In summary

It is often convenient to recognise 3 main stages in tree-breeding work:

• Selection – choosing and archiving potential breeding material
• Testing – validating the actual genetic worth of individual selections
• Production – making the products of tree breeding available to the user in commercial quantities

Under the breeding scheme described above, the breeder will reach an upper limit of genetic gain for a given level of selection intensity in this first generation. The methods exploit the genetic expression of a trait which can be passed on to the next generation in which it will become fixed. This is known as additive genetic variation. Specific crosses involving two known parents (yielding full-siblings or a full-sib family) may show genetic superiority in excess of the expectation from the individual parents which is due to non-additive genetic variation. This type of material is amenable to vegetative propagation based on family mixtures. Beyond this, more improvement can only be made by releasing more genetic variation through recombination of genes in the second generation. This will be maximised by creating families using the very best first generation parents. This type of work has now begun in Sitka spruce.