FRANÇOIS MERGEN
Assistant Professor of Forest Genetics
Yale University School of Forestry, New Haven, Connecticut, U.S.A.In preparation for the World Seed Year, FAO is seeking to assist national campaigns aimed at improving seed quality and promoting wider use of such improved seeds. An important element in the production of better seed and superior trees is research relating to genetics, and particularly to forest tree improvement. The subject of forest tree breeding research is being treated by Professor Mergen in two parts. In the present issue, he deals with individuals tree selection, progeny and clonal tests, and vegetative propagation. In the next issue of Unasylva a further article will deal with hybridization, controlled pollination, and induction of flowers and fruits. Forest tree breeding research is potentially of the greatest significance to foresters and forest products industries.
Interested readers are invited to send their comments to the Director, Forestry and Forest Products Division, FAO.
Tree planting programs are being expanded throughout the world. The demand for forest tree seed and for planting stock has increased rapidly, and frequently exceeds the supply. Under such conditions the nurseryman does not have as much control as is desirable over the genetic quality and origin of the seed. Nevertheless, very much more serious consideration than in the past is now being given to what is planted.
To alleviate the seed-supply problem, and to assure seed from trees of known quality and origin, seed orchards and seed production areas are being established. In these forest stands, selected trees are favored and propagated for the purpose of producing high-quality seed. Trees in the seed orchards are selected for a variety of different desirable traits, such as high degree of resistance to a particular disease, rapid growth, good stem form, good natural pruning, wood of high density, or high yield of oleoresin.
For seed production areas the best natural or planted stands are cleared of the inferior trees, and the desirable trees are maintained for seed production. These potential seed trees are selected purely on the basis of general outward appearance (phenotype), and the true genetic makeup (genotype) of these selected seed trees cannot be determined until the planting originating from these trees are old enough for evaluation. Tree selection for seed production areas is fairly simple and consists of leaving the best 10 to 50 trees per acre evenly distributed through the stand, so that each selected tree can attain maximum crown development and fruitfulness.
As mentioned, selection is purely visual, but if carried out in even-aged and well-stocked stands, the selected trees will normally be those that have survived competition and were best fitted for that particular site. This type of selection will probably not upgrade the resulting plantations much above their parents, but will maintain a genetic status quo in the resulting plantations, and will assure seed of known origin from good phenotypes at a cost competitive with that from commercial sources. Above all, it will tend to counterbalance the dysgenic practices of collecting cones from isolated, deformed trees; for the characteristics that make a tree attractive to the commercial collector (extreme limbiness and a large crown, along with poor natural pruning) make the tree an undesirable seed source from a forest manager's point of view.
Within the general field of individual tree selection, however, the greatest efforts have been spent not on selecting the best trees within a given stand, but within a particular locality or within its entire range. Emphasis is placed on the selection of individual trees for use in seed orchards. This type of selection is the basis or foundation for the majority of tree-breeding programs that have as a goal the mass production of trees that are outstanding in a particular characteristic. These selected "plus" or "superior" trees supply seed and pollen for progeny tests, and cuttings for vegetative propagation. They also will serve another useful purpose, namely that of preserving outstanding germ plasm for future generations. Through the forests all over the world, foresters are looking for pines with resistance to insects or disease, spruces with exceptional growth characteristics, maple trees with a sweeter sap, chestnut trees with ability to withstand the chestnut blight, or birches with curly grain, to name just a few from a large list of objectives (Figure 1).
Courtesy, Texas Forest Service, College station, Texas, U.S.A.
Courtesy, M. Lemoine, Nancy, France
Although selection of outstanding forest trees has been under way for some time, there is as yet little precise statistical information available on which to base general predictions. The outcome or benefits of a particular type of selection need to be evaluated for each case. In one instance, selection of outstanding trees along with their subsequent mass propagation might be a fruitful line of approach while, in other instances, where the desirable traits are under loose genetic control or where legitimate selection and evaluation cannot be made, this approach will yield no beneficial results at all. The total amount of improvement in a particular quality, and the rate at which the improvement will become available by selection will depend on various factors, such as the intensity of the selection procedure, the amount of genetic variation present in a particular species, the degree of heritability of the character under selection, and on the method of propagation.
Intensity of selection refers to the standards set up before one starts to look for plus trees. Selection can be on the basis of percentage of superiority, e.g., 100 percent faster growth rate than the average tree of the same stand; the best tree in a group of 100, 1,000, or 10,000 trees; or the tree that deviates by 4 standard deviations from the mean. In selection for disease or insect resistance, a choice is made of trees not affected or trees affected which were able to throw off the attack.
Selection is effective only within the range of genetic variation that is present in a species. Because in the process of selection and intraspecific controlled pollinations, no new genes are introduced into the gene pool of the species, we must have the required genetic variability to begin with, lest selection yield no beneficial results (Figure 2).
Courtesy, U.S. Forest Service
Heritability is an index of the transmissibility of the selected characteristic, and expressed in statistical terms it is the effect of the genes in relation to the total amount of variability. Therefore, the greater the genetic effect, the larger the index of transmissibility. Because heritability consists of two components - genetic and environmental - it can be modified by changing the environment.
The method of propagation plays a very important part in determining the rate at which improvement can be achieved. If the desirable traits are under rigid control, and the tree lends itself to economical multiplication by the rooting of cuttings, superior planting stock can be made available immediately on a limited scale. If, at the other extreme, the selected tree has to be reproduced sexually when mature and the rate of genetic improvement per generation is slow, a hundred or more years might elapse before practical benefits are obtained.
Rigid standards for selecting superior trees, based on the progeny data on hand, should be developed without further delay for all the commercial species so that their genetic potentials can be fully utilized. The standards and methods for selection should be worked out by joint co-operation between forest managers, wood users and forest geneticists. These standards will need to be modified as the results from replicated progeny tests become available. In these specifications individual traits should be stressed, such as high yield of extractives, rapid growth, or resistance to insects or disease. Some foresters also place emphasis on a combination of desirable traits, e.g., exceptional height and diameter growth along with straight grain, dense wood, and high alpha-cellulose content. If trees are selected for only one desirable trait, selection is most effective; the degree of difficulty and chances of failure increase in a geometric progression with the number of traits used, unless there is genetic linkage between these traits. Under most conditions, selection will be effective only if practiced on one or possibly two characters at a time.
Field selection can be facilitated by providing foresters and woodsmen with illustrated guides for the various species so that they may be on the look-out for exceptional trees. The theoretical "plus" trees can also be described in statistical terms as a basis or guide for their recognition.
For some traits, selection must be at rotation age, while for traits that "mature" in seedlings or young saplings, the "plus" trees can be selected in nursery beds or young plantations. For example, longleaf pine with a short (one- to two-year) grass stage, or trees with exceptional cold hardiness, have been selected in the nursery beds. Selection in nursery beds, however, is only of value if the characteristics for which the seedlings were selected in the juvenile stage maintain their attributes through to maturity. If selection in the nursery is feasible, advantage should be taken of this method, for it allows selection amongst millions of even-aged seedlings growing under uniform conditions. Nursery bed selection is more effective when the seed is graded by size before sowing, for this eliminates part of the variation in initial height growth and vigor.
Selection of superior trees definitely has a place in forest tree improvement programs, as has been shown by the results of F1 progeny tests from parents selected for oleoresin yield, fibre characteristics, production of rubber, resistance to drought, resistance to disease and over-all pattern of growth.
However, for most traits, as long as selection is based strictly on phenotypic or outward appearance, the ratio of "plus" genotypes to selected "plus" phenotypes will be low. Accurate diagnostic methods of selection are needed that will aid selection on the basis of morphology. From the progeny tests, methods will be evolved that use biochemical and refined physical tests of those characteristics that are the underlying causes of the desirable traits. Examples of some tests that might be used are: the tannin content of bark in chestnut as an indication of resistance to the chestnut blight, viscosity of oleoresin as an index of gum yielding ability, or resin on pine branches as an indication of resistance to the resin midge. The development of these tests should not be considered as an end in itself, but they should help to speed progeny tests for they will allow one to use very young plants.
In general, the intensity of selection should depend on what the tree is to be used for, and on how much time and effort (money) one contemplates spending on it. If the tree is to be used for extensive vegetative propagation along with controlled pollinations and subsequent progeny testing, considerable time and effort need to be invested in the selection phase. It is advisable to mark, describe and catalogue each individual tree so that this information may be exchanged between the various tree breeders, and also to protect the trees from the axe.
The number of catalogued "plus" trees that have been selected throughout the world is quite impressive. In Sweden alone, there are about 3,000 catalogued trees that were selected for their outstanding height and diameter growth, and exceptional growth and good form (Figure 3). These trees form the core of the Swedish tree improvement program (Streyffert, 1958). In Finland, by 1953, approximately 500 pines, 200 spruces, 100 birches along with some 200 trees of other species had been selected, measured in detail, and rendered identifiable in the field. Some of these selected trees are located in forest stands north of the Arctic Circle (Anonymous, 1953).
Courtesy, Dr. Arnborg, Uppsala, Sweden
When specifications for superior trees are prepared, one should keep in mind that the results of a program of controlled breeding will not be available on a commercial basis until after some 30 to 60 years have passed. With our rapid changes in utilization standards and conversion practices, some of the specific tree requirements will have been taken care of by technological advances by the time the "improved" trees are available. Therefore, one should select for traits that can still be expected to be of importance some 30 to 50 years hence.
Superior tree selection in the field is based on the outward appearance or phenotype of the tree. A tree as it stands in the forest - the phenotype - is the expression of the interaction of its genetic potential, or genotype, with that particular environment. The superior characteristics of a selected tree can, therefore, be due to either a favorable environment, or to a desirable genotype, or most likely to a combination of these two.
Without any further testing, it is not possible to ascribe a particular characteristic as being determined by hereditary factors, or by environmental forces. For example, a blister-rust-free mature eastern white pine tree can be either the result of no infection by the spores, or the tree might be resistant to the disease. Similarly, a Scots pine tree with a clear straight trunk might have shed its branches as part of its normal growth habit (genotype), or at one time it might have grown under heavy competition which resulted in a clean, straight bole. The only possible way to isolate the effect of the genotype and that of the environment is either to multiply the tree asexually and observe the propagules under a uniform environment, or to study its offspring (progeny).
Occasionally, nature makes small progeny tests for us. For instance, in an abandoned field on the Osceola National Forest, in Florida, U.S.A., the natural regeneration under an isolated slash pine that has a distinct forking habit, starts to fork at the same height from the ground and is similar in over-all appearance to the isolated parent tree. Such occurrences in nature point towards genetic differences, but controlled progeny studies are needed to obtain definite evidence that the selected characteristics, or tendencies for these characteristics, are under strong genetic control. The progeny test, although of a time-consuming and expensive nature, is essential and it will determine the amount of economic benefit to be derived from the multiplication of the selected trees. These tests form the backbone of tree improvement programs and, therefore, should be planned expertly, and executed with care. Unless the description of a superior tree is backed up by the evidence from progeny tests, these descriptions will have no use in tree improvement work, for they merely outline what a desirable tree looks like.
The rigidity of genetic control of a selected trait can be assessed by both clonal outplantings and by progeny tests. For the clonal tests, the selected trees are propagated vegetatively, and these propagules are then planted out under uniform or known environmental conditions. Propagation may be by rooting cuttings, air-layering, rooting needle fascicles, root suckers, or by grafting. Grafting, however, is not satisfactory if adaptability to a particular soil, root growth, and other related characteristics are being evaluated.
Selection of superior forest trees is generally carried out over large areas, and sometimes selected trees within a species are separated by a distance of several thousand miles. This makes it impossible to evaluate objectively the relative merits of each tree. However, by propagating these trees vegetatively, they can be brought together and outplanted in test plantations on various sites for comparative purposes.
These clonal tests, as they are called, have been used to determine if a tree is actually resistant to a certain disease or has escaped damage merely because the disease-causing organism was absent; to establish the yield and quality of oleoresin of slash pine for nave] stores purposes; to evaluate anatomical characteristics and growth traits such as needle color, branching angle! or over-all phenology. Such tests help to determine the dependence of the selected traits on the environment, but shed little light on the mode of inheritance or combining ability. They may show results quicker than progeny testing, however, inasmuch as they a low one to postulate the genotype, and obvious failures can thus he eliminated before their progeny is tested.
Dr. Larsen in Denmark was the pioneer in the use of vegetatively propagated trees as a means of comparing the genotypes of selected trees. In these "tree-shows," as Dr. Larsen calls the clonal tests, expression is sought of a uniform environment on several genotypes (Larsen, 1956). However, the only possible way to learn the mode of inheritance, or obtain an estimate of the degree of heritability of certain traits, is by studying the progeny from the selected trees. These progeny tests reveal more information on the genotype of the selected trees with reference to their breeding value than do the original trees themselves. For use in commercial plantations certain traits need to be known and the information on some of these cannot be obtained from a single tree; for example, yield per hectare.
Depending on the objectives of a tree improvement program, and depending on the manpower and physical facilities available, progeny tests with seed consists of either "one-parent" tests or "two-parent" tests. For the one-parent tests, open or wind-pollinated seed is collected and the female parent is the only one specifically known. In the two-parent test the pollen source has been controlled (by cross-pollination or self-ings) and both parents are known. Considerable in formation can be obtained on the inheritance pattern by the "one-parent tests," but the tests may need to be repeated over a period of several years, because the male parents (pollen supply) often change from year to year on account of the climatic conditions prevailing at the time of pollination. These tests, although less refined than the two-parent tests, can be started without any delay, but their evaluation will require the skill and time of trained tree breeders. They are especially useful if information is desired on how much improvement in a certain characteristic is obtainable by collecting wind-pollinated seed from superior trees.
The field designs for the progeny tests, especially the two-parent tests, present considerable problems, many of which are of a paradoxical nature (Figure 4). The number of trees in each progeny group is important, for it must be great enough to keep the sampling error down. When a large number of trees is used in each block the land area to be used is increased, which in turn increases the degree of heterogeneity in the environment. In addition, the traits under investigation partially determine the error term and thus influence directly the size of the progeny The amount of variability between the different progeny groups is mostly not uniform and this introduces another variable. In addition, one often has to deal with progeny groups of unequal number on account of failures in seed-set or loss to rodents and birds. This difficulty can be partially overcome by balanced, or partially balanced, incomplete block designs. The number of trees per hectare throughout the test will also vary, starting with a high number at the beginning and ending up with a smaller number as the trees grow larger. Depending on the method of thinning, such as systematic versus nonsystematic thinning, the error term can be considerably changed. Another consideration, not always satisfied, is that the environment has an additive effect on the genotype, and that this effect is of equal magnitude on each genotype. Some progeny tests reveal a considerable amount of information on the performance of the various progeny groups, without necessarily giving adequate information on the selected parents. Tests should be so designed that maximum correlation is obtained between the performance of the progeny and the genotype of the parents (Strand, 1952). Unless adequate replication and complete randomization is possible, it is always best to check the design with a well qualified statistician before planting the trees in the field, rather than at the time when evaluation is to be made.
Courtesy, Ida Cason Calloway Foundation, Pine Mountain, Georgia, and U.S. Forest Service
With the advance and interest in controlled environment rooms, speed-up flowering and progeny tests under rigid environmental control will permit more rapid and precise testing for certain characteristics, especially when these tests can employ biochemical or physical techniques rather than over-all growth measurements. Because the field of progeny testing of forest trees is fairly new, considerable progress could probably be made by pooling the knowledge of the geneticists, the physiologists, and the biometricians.
The technique of vegetative propagation has a direct application in forest tree improvement programs, because it allows genetically superior trees to be multiplied, 1,000 or even 10,000 times, without altering their genotype. When trees are reproduced vegetatively, they are multiplied without fusion of the male and female reproductive cells. This asexual reproduction produces plants (propagules) true to form without any different or new traits from a second parent, unless mutations occur. Thus, vegetative propagation is an essential tool in multiplying and preserving valuable germ plasm, and in estimating the genotype in clonal tests.
If a superior tree is easily rooted, it can be increased many-fold without delay. Some of the hardwoods, such as willow, poplar, or locust, are good examples of this. With most species, however, and especially the conifers where vegetative propagation on a large scale is economically impossible, multiplication has to be by seed. For this purpose seed orchards are being established with grafted stock or rooted cuttings from superior trees, and these orchards are then managed solely for the production of high quality seed.
Although of great silvicultural importance, determined attempts to propagate vegetatively forest trees that do not normally root or sprout were only started some 20 to 30 years ago. Considerable research is being conducted on finding methods to root cuttings of species refractory to rooting. Tests have been made with numerous types of media, varied temperature and light regimes, different treatments with hormones, nutrients, fungicides and insecticides, and throughout the different seasons of the year from trees of various ages (Mergen, 1955). Despite these concentrated efforts, there are still many species where it is extremely difficult to obtain cuttings with adequate root systems. With the advent of thin plastic sheets it has been possible to induce rooting by air-layering branches on trees that do not root well from cuttings. The air-layering technique, although over 2,000 years old, has only recently been used to propagate forest trees for use in improvement work (Figure 5). Certain conifers can now also be propagated by rooting individual needle fascicles.
Courtesy for Figure 5 A. U.S. Forest Service
In agriculture, various plants have been propagated vegetatively through many centuries without much apparent loss in vigor, and so far there is no likelihood of a reduction vigor in trees propagated vegetatively. There is, however, danger that because all the members of a clone are genetically identical, a disease infection or insect attack will spread rapidly once it is established. There is evidence of this disadvantage in some of the hybrid poplar clones that have been propagated by cuttings, especially where an entire stand consists of one single clone.
Depending on the species there is no or little difference in the growth pattern of cuttings taken from different parts of the same tree, and once they become established the rooted branches appear similar to trees that grew from seed. There is strong evidence, (however, that in some species cuttings from lateral branches will always have a plagiotropic growth. Araucaria provides a good example of this type of behavior.
Perhaps the greatest use of vegetative propagation lie:; not in its use as a direct means of multiplying superior trees for commercial use, but as a tool in facilitating studies. As mentioned above, selected trees are mostly scattered over the range of a species which makes it difficult to control-pollinate a large number of trees within a season. Scions (branches) from these trees are therefore grafted on stock planted close to the laboratories to facilitate the breeding work. By, this method germ plasm of the selected trees is also preserved should the parent trees die, and scions from one selected tree can be established in many locations throughout the world. There are, for instance, Douglas firs growing in Denmark that have the identical genotype a'; trees in the natural range in the U.S.A. (Larsen, 1956).
In addition to these considerations, vegetative propagation permits one to transplant a scion onto a root stock that can survive under particular conditions; change a dioecious tree into a monoecious tree; propagate bud mutants; induce flower formation; and study virus transmittal and resistance mechanisms. Inter-generic and inter-specific grafts (heteroplastic) are also possible, and these afford a great opportunity to study the translocation of solutes (Figure 6). Not all graft combinations within the same genus, however, are completely compatible, but this incompatibility is being used to good advantage with several fruit trees to induce earlier flowering.
FIGURE 6. - Heteroplastic grafting of Pinus elliottii on Picea excelsa.
A. View of plant after they were grafted for one year.B. Cross -section through graft union above, pine; below, spruce.
Great advances have been made in the general field of vegetative propagation of forest trees, but intensive studies on the physiology of root formation are needed to solve the problem of those trees that root with great difficulty. The method of vegetative propagation has proved its usefulness in various phases of forest tree improvement work, and, with the expansion of basic research with forest trees, its usefulness in other branches of forest biology will be recognized.
Forest tree breeding is still in the early stages of development - in the descriptive phase of the forming process. During the past few decades, considerable knowledge has been gained on the genetic variability within our tree species, on their physiology, on how to test the inherent variability, and on how to use this information advantageously in silviculture. More work, both basic and applied, is necessary to advance this field and, during this period and as new information is obtained, many current ideas on tree selection and progeny testing will be modified and changed. It is however, a field that can look forward to contributing greatly to the over-all practice of forest management.
ANONYMOUS, 1953. Forest tree breeding in Finland. Finnish Paper and Timber 6: 73-74.
LARSEN, C. S. 1956. Genetics in Silviculture. Oliver and Boyd, London. 224 pp.
MERGEN, F. 1955. Vegetative Propagation of Slash Pine. U.S.D.A.; Southeastern For. Exp. Sta. Paper 54. 63 pp.
STRAND, L. 1952. Progeny tests with forest trees. Hereditas 38: 152-162.
STREYFFERT, T. 1958. Forestry in Sweden. Oregon State College Press, Corvallis, Oregon. 55 pp.
The author wishes to express sincere thanks for the cooperation received from the various agencies that supplied illustrations for this article. The manuscript was critically reviewed by Professor H. J. Lutz of the Yale School of Forestry, New Haven, Conn.; Professor J. W. Wright of the Department of Forestry, Michigan State University, East Lansing, Mich.; and Dr. E. B. Snyder of the Southern Institute of Forest Genetics, U.S. Forest Service, Gulfport, Miss. Their suggestions were most helpful and are hereby acknowledged with thanks.
