J. MADDERN HARRISJ. MADDERN HARRIS is in charge of wood quality research in the Forest Products Branch of the Now Zealand Forest Research Institute at Rotorua.
MANY RECENT reviews are available which cover almost all aspects of breeding trees to improve wood properties. Little is to be gained from attempting to include the entire literature of this field once again. The aim of this paper is to summarize recent knowledge, to review practical advantages of wood quality improvement, and to indicate priorities for application of such improvements and for future research.
The term wood quality will be used to refer to the properties of clearwood only. The causes and consequences of defects such as knots, reaction wood, grain deviations and most other forms of xylem discontinuity belong logically to the section of this consultation dealing with stem quality. Nevertheless, it is necessary to stress that superior wood quality can often become fully effective only in the presence of superior stem quality. For example, the natural variability within many softwoods results in the strongest piece of defect-free wood produced by a particular species being perhaps two or three times as strong as the weakest piece; yet, the presence of defects can reduce the strength of either by a factor of 10 or more. A large cluster of encased knots or severe sloping grain in a critical region can render wood of the highest intrinsic quality mechanically useless. Consequently, any prediction of the gains to be derived from improved wood quality will, for most uses, require the added proviso, other circumstances being equal. The most important circumstances will be those affected by stem quality.
As stem characteristics have proved deserving of and amenable to improvement by breeding, it is probable that they will frequently be accorded priority over wood quality in breeding programmes. On the other hand many wood products, especially board and paper products, are very sensitive to variations in wood quality, and no opportunity for improvement, should be overlooked.
General summaries of the heritability of v-god properties have been prepared by Goggans (1961), Hattemer (1964) and Zobel (1961 and 1964). The heritability of wood density has been summarized by Harris (1965) and Zobel (1965a), and work on tire heritability of fibre properties has been covered by Smith (1965), van Buijtenen (1965) and Zobel (1965a). Zobel (1965b) has also reviewed the inheritance of spiral grain.
Any brief assessment of the very considerable body of work covered by these summaries must recognize the following features:
1. The great bulk of the work is very recent, most of it having been carried out over the past 10 to 15 years. Consequently, workers nave been feeling their way, developing new techniques both of assessment and of analysis. They have often been obliged to use whatever material was available rather than being able to survey some ideal array.2. Because the research field is relatively new, most of the results refer to young trees rather than to trees of merchantable age.
3. At best an estimate of heritability represents the specific result of a well-designed experiment carried out under clearly stated conditions. An estimate should never be regarded as an absolute or unalterable property of a species. To be of maximum value, the estimate should be based on adequate data obtained from contrasted sites. It should be presented together with information on the environmental conditions, the sort of population used, and the time intervals between observations. Few of the results so far reported achieve all these ideals, and fewer still can be used to assess genetic gains.
Advantages of using computed expected genetic gain as the basis for comparing improvements in wood properties are well illustrated by the tabulated values presented by Namkoong et al. (1967). Unfortunately most published data are not presented in this form and are not suitable for computing genetic gain. For this reason evidence will have to be discussed in terms of heritability estimates in this paper.
For most of the wood properties examined so far, the results carry conviction of the role of genetics from the cumulative evidence rather than from the strength of results of individual experiments. Nevertheless, some general trends are beginning to emerge. and a number of specific results do hold considerable promise for future gains. It will be convenient to discuss each wood property separately. First the applications of each property will be outlined (see also Dinwoodie, 1965; Nicholls, 1967) and then the available evidence for improvement of each property by breeding will be examined.
WOOD DENSITY
This property is closely correlated with major strength properties of wood, with pulping yields, and with pulp and paper quality, as well as with machining, gluing and finishing characteristics. For these reasons, wood density is the property that has been most widely studied genetically, even though it is a complex property dependent on a wide variety of anatomical variables (Nylinder, 1965). Despite this complexity, fairly large narrow- and broad-sense heritabilities have been reported for coniferous trees 5 to 15 years old. Heritability values for corewood (1 to 5 growth layers from the pith) are usually lower. Relatively little work has been done on hardwoods, and- results so far are less promising. Broad-sense heritabilities of 0.2 to 0.4 (for example, for Populus spp.) are typical. Some reports of heritabilities much higher than this imply that much more work is justified before the possibility of improving hardwood densities can be dismissed.
FIBRE PROPERTIES
For convenience the term fibre will be used in this paper to refer to true fibres of hardwoods and also to softwood tracheids. Variations in fibre properties are the main determinants of variations in wood quality. The proportion of fibres to other types of cells in hardwoods and their dimensions in hardwoods and softwoods influence wood density, and they also largely determine the quality of pulp and paper products (Dinwoodie, 1965).
Fibre length
Inheritance studies of fibre properties are still in the early stages, but it appears that the degree of genetic control of fibre length is at least as strong in conifers as that of wood density. Known inheritance values for hardwoods are few, but some rather high values have been assumed from related data.
If these assumptions are true, improvement of fibre length in hardwoods may repay further study. Though the range of values is limited (Zobel 1965a), fibre length is such a limiting feature to utilization of hardwoods that even modest gains would be welcome.
Fibre diameter and wall thickness
Considerable variation has been recorded in these properties in conifers, but very few inheritance studies have been published. Results so far have provided good evidence of genetic control, but it does not appear to be as strong as control over fibre length.
The strong heritability values obtained for wood density are largely dependent on these fibre characteristics. It therefore seems that the large experimental error that is possible in measuring diameter and wall thickness may have depressed estimates of their heritability.
Proportions of fibres in hardwoods
It is well known that the proportion of fibres to vessels and parenchyma cells in hardwood is subject to environmental control, but little information is available on the importance of genetic control except from clonal studies of Populus hybrids.
SPIRAL GRAIN
The importance attached to spiral grain varies widely from country to country. Attitudes probably depend on the severity of its occurrence in locally used species, on end uses, and on the climatic contrasts which focus attention on the dimensional instability of spirally grained timber. Where problems do arise from spiral. grain, they can be severe indeed.
Very few of the numerous published results contain statistical estimates of the heritability of spiral grain. Heritability has generally been assumed to be high. This is borne out by Nicholls et al. (1964) who found narrow-sense heritability to be 0.66 for corewood of Pinus radiata, with lower values outside the corewood. Spiral grain is most marked in the corewood of this species, but it is not normally an important feature of outer wood. On the other hand, Zobel et al. (1968) found genetic variation in P. taeda to be of a nonadditive type, so recurrent selection to reduce spirality would not be very effective in this species. They also noted that the incidence of spirality does not appear to be widespread or serious in P. taeda.
CHEMICAL PROPERTIES
Extractives
The heritability of extractives contained in wood (as distinct from chemicals obtained by tapping living trees) has mainly been examined in the softwoods used for pulp and paper manufacture. A high content of extractives can reduce yields, increase the consumption of chemicals in processing and may severely limit the usefulness of a timber for groundwood pulping. Though the content of extractives differs widely between trees, studies so far have not indicated strong heritability (van Buijtenen, 1967). The influence of extractives on wood utilization is very extensive (Hillis, 1965), and the possibility of modifying their effects by breeding deserves much more attention in future.
Heartwood formation
An assessment of heritability in heartwood formation, though of great importance for many wood uses, requires, in most species, older trees than are generally available as clonal or family assemblages. The only published results are those of Nicholls (1965) in which gross heritabilities of 0.36 are reported for heartwood formation in P. radiata.
Cellulose/lignin ratios
Cellulose content could be important in determining the yields of chemical pulps, but the few studies carried out so far [for example, Zobel et al. (1966) on P. taeda] indicate such a small degree of inheritance that selective breeding would give little improvement.
The list of wood properties studied so far is by no means exhaustive. For example a recent paper by Zobel et al. (1968) has underlined the importance of moisture content in wood used for pulping. Variability of moisture content within a species may be primarily a function of variation in wood density and heartwood formation, but further study is undoubtedly needed on this topic. Similarly, brightness is a critical feature of many softwood pulps, but fluctuations in brightness have yet to be related to variations in the raw material, even though intraspecific variation is quite widely regarded by pulping operators as the cause.
Perhaps the current situation can best be summed up by saying that many important wood properties can undoubtedly be improved by tree breeding. This improvement need not be given absolute priority. If necessary it may be claimed as a bonus in other improvement programmes. Thus, wood properties could be considered during final selection of parents initially chosen for other reasons, or during later additions to a programme.
An important point that emerges is the need to judge each species, or even each strain within a species, on its own merits. No generalization can be assumed. We know too little about the ultimate physiological bases of variations in wood properties to assume that, because a particular wood property does not prove amenable to improvement by breeding in one species, similar results will be obtained even in closely related species.
Consequently, the greatest need at the moment is for published results of soundly based experiments to examine genetic variances and also to measure actual genetic gains. Large-scale, adequately planned genetic experiments, along the lines indicated by Bannister (1964) and Stonecypher (1966), are likely to be few in number. Therefore, for many years we shall have to rely on cumulative evidence from progeny tests aimed at modifying one or two properties of real importance.
Tree breeding is only one of several processes that can be employed to manipulate wood properties. As these will form the background to any breeding programme, they must be mentioned briefly at this point. One example will suffice to indicate the general role of each process.
FOREST MANAGEMENT
Management practice regulates wood properties in many ways - through species selection, through changing growing conditions as determined by the sites selected for each species, and through many other decisions such as those regulating the ages of rotations (van Buijtenen, 1969).
Example
Long rotations can be used to increase the proportion of sawlogs furnished to pulp mills in integrated operations. In general, long rotations will produce a high proportion of outer wood. For most softwoods, the resulting pulp chips would have greater wood density, longer fibres, thicker cell walls, more heartwood and extractives.
SILVICULTURE
Silviculture controls tree growth through initial spacing, thinning and site amelioration. By these means rates of growth can be manipulated within broad limits. In some species, wood properties such as density and percentage of latewood can be modified correspondingly.
Example
Softwoods grown in open stands tend to form less latewood than when grown under dense stand conditions (Larson, 1969).
WOOD SELECTION AT THE MILL
In a small sawmill there may be little need for wood selection beyond an ability to sort logs by species and to optimize the cut from each. In large integrated mills producing sawn timber, veneers, particle board and/or pulp and paper products, wood selection can be very complicated.
Example
Pruned logs must be separated from unpruned logs. Logs from tops of trees may be chipped for chemical pulp or particle board, or they may be diverted to the groundwood mill. Sawmill slabs may be used to furnish a long - fibred admixture to chips from other sources. The variety of wood properties that becomes available when several species are used is limited mainly by the demands of the different end uses for sufficient material to meet the requirements of production. In this sense the various sections of an integrated operation become competitors and only advanced systems analysis and operations research can decide what is the most economical division of the available timber between them.
Put simply in this way, it might appear that so many other options are available to timber producers and users that wood improvement by breeding is an unnecessarily laborious and time - consuming process, and one, that is hardly worth the effort. This might be so if the economics of management and silviculture did not place severe restrictions on what can reasonably be done in the field and if many end uses did not demand the same sorts of high quality material. As it is, economic limitations usually prevail, and many processes that start off as waste users eventually demand wood of better quality to meet marketing competition.
The extent to which tree breeding is used to improve wood properties required for a particular product must almost always be decided by economic considerations. The data required for economic analysis are:
1. The need for improvement projected in terms of technological trends over the next one-and-a-half stand rotations - that is, looking ahead not less than 30 years, even for fast-growing conifers.2. The degree of improvement which is required and feasible. For example, are wood fibres 3.4 mm long really superior to fibres 3.2 mm long, and what would be gained if we aimed at 4 mm ?
3. The cost of achieving such improvements by breeding. Visual selection for stem quality will almost always come first, followed by the more difficult laboratory evaluation and reselection for wood properties. If improvement in more than one property is sought, compromises must be made in selection intensity and projected gain for each. If desirable characteristics are negatively correlated, the cost of remedying losses in one property must be debited against the gain from the other.
4. The increased value that various degrees of improvement will confer on the product.
5. The costs and benefits of the breeding programme must then be balanced against the cost and benefits of alternative methods if any are applicable.
It must be conceded that the value of improvement may not always be directly assessable in monetary terms. For example, if quality improvements are required to meet competition, there may be no direct cash benefit, but the industry may fail if no improvements are made. However, this only emphasizes the need to assess wood quality improvement (by whatever means it is obtained) as part of the cost of production.
Much of the information required for economic analysis is now available or will become available as breeding programmes come to maturity. Unfortunately there is a dearth of published data on the values of qualitative improvements in various. products. This and the shortage of reliable estimates of genetic gain for many wood properties comprise the two most immediate needs for assessments of breeding programmes. A pioneering attempt by Namkoong et al. (1967) evaluates the results of increasing wood density when P. taeda is used for chemical pulping and fully acknowledges limitations imposed by data that permit analysis only in terms of total yield of cellulose. They conclude that: "based on these data, which are admittedly inadequate, one could rationally say that the best breeding procedure would be to give specific gravity a negative weight or at best ignore it in breeding programmes." This controversial conclusion should stimulate others to provide data for similar analyses that will include the corresponding qualitative gains.
Because of the rapidly changing demands of technology, improvements of only fundamental wood properties are likely to prove worth while. For many species other selection criteria have priority; wood properties should be included primarily to ensure that they are not affected adversely. Alternatively, the main objective may be to reduce within-tree variability in some fundamental property; this would contribute to easier and more efficient harvesting and manufacture and could ultimately prove to be the greatest contribution of wood quality improvement. But properties fundamental to one industry may be of little concern to another.
The most clear-cut case for wood quality improvement would arise in an industry producing a single end product from a tree species of such outstanding merit that only vigorous stems of good form were produced and in which all wood properties but one fully satisfied the needs of the product. If the deficient property proved to be more economically corrected by selective breeding than by any other means, and if the cost could be justified by the increased value of the product, then a watertight case for improvement would be established.
This hypothetical case may appear to be pure satire, or at least a sophistry, but the reason for advancing it is to emphasize the absolute interdependence of all criteria for tree improvement. It should be axiomatic that any change in management or silviculture or any change or diversification in end products must inevitably have repercussions on breeding strategy. Breeding programmes must be flexible and adaptable. These in turn will require flexible methods of assessment. The " four tree tester " design (Zobel and Kellison, 1963) is an example of this approach to meet the needs of a particular programme.
BANNISTER, M. H. 1964. Planning a genetical survey of Pinus radiata populations. FAO/FORGEN 1:4/1.
DINWOODIE, J. M. 1965. The relationship between fibre morphology and paper properties. Tappi, 48(8): 440-447.
GOGGANS, J. F. 1961. The interplay of environment and heredity as factors controlling wood properties in conifers with special emphasis on their effects on specific gravity. North Carolina State College, School of Forestry. Tech. Rep. 11.
HARRIS, J. M. 1965. The heritability of wood density. Proc. IUFRO Section 41. Vol. 2.
HATTEMER, H. H. 1964. Estimates of heritability published in forest tree breeding research. FAO/FORGEN 1: 2a/3.
HILLIS, W. E. 1965. The influence of extractives on wood utilisation. Proc. IUFRO Section 41. Vol. 1.
LARSON, P. R. 1969. Wood formation and the concept of wood quality. Yale Univ., School of Forestry. Bull. 74.
NAMKOONG, G., BAREFOOT, A, C. & HITCHINGS, R. G. 1967. Proc. 4th Tappi Forest Biology Conf.
NICHOLLS, J. W. P. 1965. Heritability of heartwood formation in Pinus radiata D. Don. Nature, 207:320.
NICHOLLS, J. W. P. 1967. Selection criteria in relation to wood properties. Proc. Australian Forest Research Group. No 1.
NICHOLLS, J. W. P., DADSWELL, H. E. & FIELDING, J. M. 1964. The heritability of wood characteristics of Pinus radiata. Silvae Genet., 13(3): 68-71.
NYLINDER, P. 1965. Non-destructive field sampling systems for determining the wood density of standing timber over large areas. Proc. IUFRO Section 41. Vol. 2.
SMITH, W. J. 1965. The heritability of fibre characteristics and its application to wood quality improvement in forest trees. Proc. IUFRO Section 41. Vol. 2.
STONECYPHER, R. W. 1966. The loblolly pine heritability study. International Paper Co., Southlands Expt. Forest. Tech. Bull. 5.
VAN BUIJTENEN, J. P. 1965. Inheritance of fibre properties in North American conifers. Proc. IUFRO Section 41. Vol. 2.
VAN BUIJTENEN, J. P. 1967. Pulpwood properties and tree breeding: a synthesis. Proc. 14th IUFRO Congress, 9: 243-62.
VAN BUIJTENEN, J. P. 1969. Controlling wood properties by forest management. Tappi, 52(2): 257-259.
ZOBEL, B. J. 1961. Inheritance of wood properties in conifers. Silvae Genet., 10(3): 65-70.
ZOBEL, B. J. 1964. Breeding for wood properties in forest trees. Unasylva, 64: 89-103.
ZOBEL, B. J. 1965a . Inheritance of fibre characteristics and specific gravity in hardwoods: a review. Proc. IUFRO Section 41. Vol. 2.
ZOBEL, B. J. 1965b . Inheritance of spiral grain. Proc. IUFRO Section 41. Vol. 1.
ZOBEL, B.J. & KELLISON, R.C. 1963. Progeny testing seed orchards. N.C. State-Industry Co-operative Tree Improvement Program. (Mimeo)
ZOBEL, B. J., STONECYPHER, R., BROWN, C. & KELLISON, R. C. 1966. Variation and inheritance of cellulose in the southern pines. Tappi, 49(9): 383-87.
ZOBEL, B. J., MATTHIAS, M., ROBERDS, J. H. & KELLISON, R. C. 1968. Moisture content of southern pine trees. North Carolina State Univ., School of Forest Resources. Tech. Rept 37.
ZOBEL, B. J., STONECYPHER, R. W. & BROWNE, C. 1968. Inheritance of spiral grain in young loblolly pine. For. Sci., 14: 376-379.
WORLD DIRECTORY OF FOREST GENETICISTS AND TREE BREEDERS
Associated with the Second World Consultation on Forest Tree Breeding was the preparation of an up-to-date World directory of forest geneticists and tree breeders.
Lists of forest geneticists and tree breeders had been published by the Society of American Foresters in 1960 and 1962 in the Journal of Forestry. These lists were extensively used by international organizations such as FAO and IUFRO, by research organizations throughout the world, and by individual researchers. By 1967 revised lists were clearly needed and the Tree Improvement Committee of the Society of American Foresters began to prepare a new directory. Dr. Hans Nienstaedt, of the Institute of Forest Genetics, Star Route 2, Rhinelander, Wisconsin 59501, a member of the Tree Improvement Committee, was responsible for the work of preparation.
More than 1400 questionnaires were sent out by mail early in 1968 and the draft of a new directory was made available for review at the Consultation in Washington. This resulted in many corrections being received and more than 300 names added. The final version contains 1134 names arranged alphabetically by countries, with addresses, subject-matter interests and general interests, together with a general alphabetical index of names, and indexes of subject matter and genera. It is hoped to issue the directory as a supplement to the proceedings of the consultation.
