BRUCE J. ZOBEL
BRUCE J. ZOBEL is E. F. Conger Professor of Forestry at the School of Forestry, North Carolina State University, Raleigh, North Carolina, United States. He was assisted in the preparation of the final text of this chapter by P. R. LARSON (United States). Other members of the drafting team were E. W. J. Phillips (United Kingdom), and G. Siren (FAO).
Discussion and definition of wood properties emphasize that the geneticist usually works with "basic" units of wood, such as the tracheid, but real improvements have been obtained when working with combinations of properties, such as specific gravity (basic density).
Ways to express genetic improvement are considered here in detail. The need for care in the use and interpretation of broad-sense and narrow-sense heritabilities is stressed. Although heritabilities, per se, are of interest, it is necessary to assess improvement in terms of genetic gain, which is dependent on both intensity of inheritance and the amount of variation present. Few studies list results in terms of gain.
Methods of reporting gains are discussed, emphasizing the need for caution in extrapolating gains observed in seedling progenies to mature trees. Inheritance patterns change with age, and one may obtain erroneous results from estimates based on juvenile performance. Narrow-sense heritabilities for specific gravity have increased rapidly up to 15 years (from about 0.25 to 0.5 or above), as has tracheid length. Some characters, such as spiral grain, show a decrease in inheritance with age.
It is essential, in reporting or assessing percentage gains to know precisely what base is used for determination of the percentage. If a figure of 10 percent improvement is reported, the question must be asked: Improvement over what? Much confusion is caused by use of different methods of reporting percentage gains.
Actual inheritance patterns and data reported in the literature are given for geographic variation in natural stands. The value and limitations of such studies are pointed out. Many geographic variation studies have been made, and in most instances considerable variation has been reported. Individual tree variation has usually been large, but this reveals nothing directly about intensity of inheritance.
The study of geographic variation in provenance tests has yielded substantial evidence of racial diversity in several wood properties. Sometimes the patterns change from the relative position occupied in the native habitat or region. Considerable evidence of inheritance of specific gravity and tracheid length associated with source of seed is presented, but the evidence available for relationships of chemical or other wood characters is much less.
Consideration of inheritance of specific gravity through sexual reproduction shows that much data are available showing that heritabilities (and yield improvements) may be quite large. Narrow-sense heritabilities have been remarkably high for the complex character specific gravity, with values reported as high as 0.6 or 0.7. These approach the values of broad-sense heritabilities for specific gravity.
The discussion on inheritance of specific gravity through vegetative reproduction is based on data concerning broad-sense heritabilities. Values are remarkably similar for several species, ranging from 0.50 to 0.85 for conifers. The values were much lower for species of Populus.
The inheritance of characters other than specific gravity is reviewed. In nearly every instance, notably with tracheid length and tracheid widths but with the major exception of cellulose yields, inheritance in both the broad sense and narrow sense is quite strong. Most reports of inheritance are in general terms, with progeny values intermediate between parental values.
Broadleaved species are discussed in this chapter, but few instances of data on narrow-sense heritabilities have been found. Values for broad-sense heritabilities of poplars show specific gravity to be low and tracheid length high.
The philosophy of use of genetics of wood in tree breeding is discussed. Usually, growth rate and form cannot and need not be sacrificed in breeding for better wood.
Variation patterns in wood are large enough and inheritance patterns strong enough to make possible real gains of practical significance the framework of better growth and form.
Environmental factors strongly modify wood qualities by their influence on the growth of the tree. The determination of heritabilities and the assessment of genetic components of wood quality require that the contributions of environment be fully recognized and measured. In considering wood quality as influenced by both heredity and environment a dynamic relationship is recognized in which each genotype responds to the changes in the environment in its own characteristic manner.
In forestry, the most important end product for which trees are usually grown is wood. Despite this clearcut and major silvicultural objective, it is surprising how much of the early forest genetic and tree improvement research has been concerned only with improvement in growth rate, habit, adaptability, pest resistance, and the like. Wood, per se, of the improved trees has often been ignored and only in the past few years has the importance of wood qualities and yields become generally recognized as an integral part of forest tree breeding programs.
A wood property, in common with other tree characters, is determined by the environmental conditions under which a given species develops, as well as by its inherent nature. Breeding for wood properties is a vast and complex subject, a complete treatment of which must consider both the environmental and genetic factors which control wood qualities. Such a complete treatment is beyond the scope of this chapter. Therefore, the present discussion will be largely limited to genetic control of wood qualities, although it is fully recognized that environment and growth factors are an integral part of this problem. In fact, the genetic and environmental effect are frequently confounded and scarcely separable.
The need for information on genetics of wood is very great. The large areas of seed orchards now being established should have as one major object the improvement of wood properties. In fact, the value of tree breeding cannot be adequately assessed until the degree of control possible for wood properties is determined and recognized. The importance of wood is so axiomatic that it is unthinkable not to strive for its betterment in improved cultivars of forest trees. Yet this is neither simple nor susceptible to rapid execution.
Several deterrents are immediately encountered when breeding for wood properties. First is the difficulty of determining the kind of wood desired. It is easy enough to see why there is confusion and difficulty in arriving at a decision as to desired wood properties. Most wood users simply do not know, except in a vague and general way, how their final product is affected by different wood qualities; thus, they have been unable to tell the geneticist what to breed for. Furthermore, no single kind of wood is ideal for every purpose. In the absence of such directives, the geneticist either has ignored wood qualities altogether or has concentrated simply on increasing yields. This raises the problem of the relative importance of wood yield versus wood quality. Nearly everyone will agree that well-formed, disease- and insect-resistant trees that produce maximum volume yields are desired. But there is no unanimity as to how much quality of wood per se should be emphasized.
Breeding for wood quality has been delayed because of the widespread belief that wood is a "conservative" part of the tree, less subject to selective forces and therefore less variable genetically than other characters of the tree. However, wood is quite variable and a significant portion of the variation stems from genetic factors. Moreover, the characteristics of wood change with the age of the tree. This change with age is complicated by a difference in the relative contribution of genetics and environment in the determination of wood qualities, making interpretation of results difficult and sometimes inaccurate or misleading.
Genetic studies of wood have also been discouraged by the great reluctance of those primarily concerned with wood research to recognize the large variation of wood qualities within trees as well as between trees of the same species growing in different geographic areas or on different sites. There has been a strong tendency to explain variation of wood only on the basis of differences in growth rate, in environmental conditions, or both. Results of wood genetic studies are only now beginning to trickle in, and most studies are so recent that results are based only on young trees. Genetic findings on young trees result in much good information which can guide applications in tree breeding, but their applicability to older trees is restricted. The conclusions in this chapter, therefore, are merely forerunners of large-scale results expected soon to be available from well-designed experiments, all of which will increase in value as the trees being observed come nearer to harvesting age.
What is a wood property?
The term "wood property" is frequently used in a loose and ill-defined manner. The property commonly referred to as specific gravity or basic density1 will serve to illustrate this point. Although one of the most useful concepts categorizing wood, specific gravity is not a character in the sense that it can be manipulated genetically, as a unit. It was clearly pointed out by Schreiner (Zobel, 1956) that specific gravity really consists of a complex of characters, each of which contributes to the overall specific gravity and each of which may or may not be genetically independent of the other. The individual wood fiber or tracheid is the "basic unit" usually employed in describing wood morphology (from a genetical standpoint). Thus, specific gravity is largely determined by the several different "units" of cell dimensions, such as width of cell, thickness of cell wall and proportions of thick- to thin-walled cells within an annual ring. Each of these "units" of specific gravity is susceptible to genetic manipulation. Although it may be technically wrong to consider specific gravity as a unit character, nevertheless good gains have been demonstrated using specific gravity as the smallest unit of measurement. It is obvious, however, that maximum efficiency in breeding for specific gravity cannot be achieved until it is known how to control the components determining it. For example, Goggans (1962) found that width of springwood tracheids and double wall thickness of summer-wood tracheids of Pinus taeda were strongly inherited, while width of summerwood tracheids and wall thickness of springwood tracheids were less strongly controlled genetically. He found that overall specific gravity of the tree and summerwood specific gravity were quite strongly controlled genetically but that specific gravity of the springwood showed the weakest genetic control of all the characters examined.
1 Basic density is defined here as the ratio of dry weight to green volume. Basic wood density is often used synonymously and in the metric system is equivalent to specific gravity.
Any discussion of wood characters would be incomplete without inclusion of chemical characters, such as cellulose, hemicellulose, lignin or extractives. These have been less well studied than those dealing with fiber morphology, but they play a most important role in the end uses of the wood, particularly because of their effects on workability, strength, resistance to insects, diseases and the like. Research in genetic control of chemical characters has scarcely scratched the surface and is greatly needed.
Measurement of genetic control
There are several ways to express the intensity of genetic control of a character. Statements that a wood character is strongly or weakly controlled genetically can be helpful but frequently also meaningless, or even misleading, unless "strong" and "weak" are defined adequately. Thus, lists of important wood characters with reference to the relative degree to which they may be influenced by genetics and environment (Schreiner, 1958; Dadswell and Wardrop, 1960) have been helpful but need cautious interpretation and can be employed only as general guides. To be most useful inheritance should be expressed in unambiguous terms, which are consistently employed. It should be determined so as to permit an estimation of the gain possible, if a certain breeding procedure is applied to a given species growing under certain defined conditions.
The most commonly employed quantitative measure of inheritance is the ratio called heritability. As will e seen from the explanation in Chapter 2, this ratio expresses the relative importance of genetics and the environment in determining a specific character. Hanson (1963) in an able discussion on heritability states: The concept of heritability is simple. Discrepancies rise when the definition is applied to breeding situations - heritability is used in both a broad and a narrow sense. For the broad sense, the genotype is considered as the unit in relation to the environment. However, genes segregate and come together in new combinations exhibiting intra-allelic interactions (dominance) and inter-allelic interactions (epistasis). The differences between the actual effects of genes in combination and the average effect in the population are dominance and epistatic effects which are transmitted only in part. Heritability in the broad sense, therefore, considers total genetic variability in relation to the phenotypic variability while heritability in the narrow sense considers only the additive portion of the genetic variability in relation to the phenotypic variability."
Confusing broad- and narrow-sense heritabilities has been common in interpreting results of wood genetic studies. A number of persons cite heritability figures without defining clearly whether they are used in the broad or narrow sense, with the consequence that some foresters have obtained a distorted idea as to the improvement in wood possible through genetic manipulation. As Toda (1961) clearly points out, ² heritabilities determined on vegetatively propagated progenies are broad-sense heritabilities, while narrow-sense heritabilities can be determined only from sexually propagated plants.
² His definitions are based on the ideas of Lush (1949).
Another source of confusion arises from the several statistical approaches to calculation of heritability. One commonly employed method is by analysis of variance among progeny groups, even though nothing may be known about the parents involved. The results obtained may vary considerably depending on whether progeny group means or individual tree values are used in the computation. Heritabilities are often reported in the literature without any indication as to whether they were calculated on progeny mean or individual tree data, without any mention of the degrees of freedom involved or without a clear description of the design of the experiment. This description is very important because the designs actually employed in progeny-testing the value of trees in breeding experiments may be faulty or inadequate for the provision of reliable estimates of variance components necessary to determine heritabilities. In addition, some confusion and controversy exist as to the most "correct" method of placing confidence limits on the heritability obtained.
Narrow-sense heritability can also be determined by the offspring-parent regression method. The statistics involved are simple but this method requires an accurate assessment of the parent-tree wood, which is often difficult or impossible to obtain. For example, if progeny are 5 years old and the parent trees 50 years old, it is not considered statistically acceptable to make a regression of the two ages of wood directly. In this case the wood of different ages must be considered as two separate characters. The problem then is to determine the genetic correlations of the two characters as well as the heritabilities of each, since progress made in selection is a function of these three values. It is always possible, of course, to obtain wood of the first 5 years from the parent, but the formation of heartwood, with its cumulative depositions of substances, incidence of rot, possible slow growth of the "wild parent tree" when young, and excess compression wood, all make direct comparisons of the same age of wood difficult. This problem is discussed in some detail by van Buijtenen (1962), who reports heritabilities calculated on the relationship of juvenile wood of the progeny to juvenile wood of the parent and juvenile wood of the progeny to mature wood of the parent. He found the latter to be much larger, but because of the age differential in wood they can be misleading. Under these conditions, van Buijtenen presents the idea that the correlation coefficient gives a more accurate estimate of heritability than the regression coefficient. He reasons that the spread of values in juvenile wood is much less than that in mature wood of the parents, and thus statistically the correlation coefficient gives the most accurate assessment of the relationship.
It is essential to recognize that in offspring-parent determinations for the inheritance of wood comparisons are rarely made between woods of the same age produced under the same condition. Despite this, some remarkably high values have been obtained when comparing mature wood of the parent with juvenile wood of the progeny, even when grown in completely different environments. For example, Smith (1962) showed a good relationship of specific gravity between mature parents and 3-year-old progenies of Pinus taeda when mature wood sampled at breast height ³ of the trees growing in Texas was compared with bolewood of the progenies grown in North Carolina over 1,000 miles from the parent source and under a completely different environment.
³ 1.35 meters or 4,5 feet above ground.
Another difficulty in assessing inheritance in wood properties involves time. Heritability of a wood property will change with age of tree, and the values obtained from young wood may not indicate what will be found later. For example, narrow-sense heritability of specific gravity may be low in young pine seedlings but becomes higher with age; at least up to 15 years for the studies reported. Such a change of heritability was shown experimentally by Stern (1958) who, in a model research project with snapdragons, found a systematic change with age. With 2-year old Betula verrucosa he found a total genetic variance of 3 percent; when the trees were 3 years old, the variance was 25 percent; in the fourth year it rose to 53 percent (Stern, 1960). However, it is by no means certain that heritability of all wood characters increases with age. Specific gravity and tracheid length appear to follow this pattern, but spiral grain might follow the opposite pattern showing a lower heritability with age of tree. In any event, it is necessary to recognize that heritabilities now being reported based on data from young trees may not be too indicative of the heritability realizable when the same trees are old enough to harvest.
The point that heritability, as such, may not be too meaningful in tree breeding is worthy of additional emphasis; some plant breeders have rightly criticized forest geneticists for excessively emphasizing heritability per se. What is needed are measures of realizable heritability or genetic gain, not in seedlings or young trees but at time of harvest.
Genetic gain and improvement possible
The tree breeder is naturally interested in the maximum expected improvement. It is frustrating to find that so much of the literature deals with results only in general terms, usually reporting statistical significance of mean differences. Even though differences may be statistically significant, they may not be of sufficient magnitude to provide a basis for improvement through breeding. Gains depend not only on the magnitude of heritability but also on that of the selection differential; as a result, simple statements of heritabilities are not too illuminating. It is understandable why only simple heritabilities have been so widely quoted; since most reported results are for young progenies, it is difficult to extrapolate them into gains at time of harvest. Nevertheless, as a guide to the forester, the geneticist should attempt to determine whether the potential changes in wood through use of genetics have any real meaning. It seems worthwhile to speculate, as did Klein (1956), who reported that 5 cubic meters of Norwegian Picea abies but only 4.5 cubic meters of German Picea abies are required to produce 1 ton of sulphite pulp. Since the German Norway spruce also grows faster, Klein estimated that it would produce 20 to 25 percent more cellulose per hectare than Norwegian Norway spruce.
Perhaps one of the greatest sources of confusion in estimating genetic gains arises from reporting them on a percentage basis. The literature includes estimates ranging from 1 to 100 percent for expected gains in wood properties. "Percentage of what?" is the question that immediately arises, and percentage estimates can be quite misleading if not accompanied by a clear definition of the base used. To emphasize this point, consider the following simplified example from studies in North Carolina, U.S.A., on Pinus taeda using some generalized assumptions for ease of illustration. It has been found that, with the moderate intensity of selection for specific gravity used in seed orchards, it is quite possible to obtain an increase of about 34 kilograms (75 pounds) in dry weight per standard cord in one generation. Assuming a cord weighs 1,180 kilograms (2,600 pounds) dry, this gain amounts to about 2.9 percent based on the indicated average cord weight for the species. However, if one wishes to determine the weight difference between a cord of the highest and a cord of the lowest specific gravity trees to be found in the species, a differential of about 454 kilograms (1,000 pounds) per cord dry weight is obtained. Then the 34 kilograms (75 pounds) gained, compared with the biologically maximum difference of 454 kilograms (1,000 pounds), represents a 7.5 percent improvement. The 2.9 percent improvement indicates an average added weight gain per cord which is of primary concern to the forester, but the 7.5 percent indicates to the geneticist the proportion of the possible biological improvement that might be obtained. Similar calculations for tracheid lengths would show improvements of approximately 11 percent and 14 percent respectively. To clarify the meaning of percentage gain, the basis of calculation should always be indicated.
A recent summary of inheritance of wood properties in conifers was made by Zobel (1961). This summary will be used as the basis for discussion of earlier studies, and the present chapter brings up-to-date information on inheritance in conifers. In addition, genetic studies of wood in angiosperms will be reported, although it is clear that there is a serious lack of such information for broadleaved species.
In early papers, inferences about genetic control of wood were often based on the variation patterns present in wood of existing wild populations of forest trees. These variation patterns can be broadly grouped into:
1. variation within trees;
2. variation between trees;
3. variation between sites;
4. variation between geographic origins.
These variation patterns in wood have been known for many years, yet many foresters have ignored them. It is remarkable how many papers reporting wood properties for a species or group of species have based conclusions on samples from one, two, or at best very few trees from one stand or from one geographic area. Occasionally little or no attention was even paid to the location of the samples within the tree. Although some current studies still ignore these sources of difference in wood, there is now a general awareness of the variation patterns in wood properties. Hundreds of recent papers have emphasized the magnitude of within and between tree variation patterns; there is no space for such a lengthy discussion to be entered upon here.
When a number of studies showed major variations in wood among trees of the same age, the same crown class, growing with their roots intertwined in the same microsite, researchers surmised that genetics must account for at least some of the observed variation and were hopeful that strong genetic control might exist. Such observed variation among similar trees is no proof of genetic control, however, and only properly designed and analyzed clonal or seed progeny tests can indicate whether there is a genetic component in the determination of wood.
The remainder of this chapter will be devoted to a discussion and analysis of available research results considered most pertinent in establishing the proof and magnitude of inheritance of wood properties.
Geographic variation - natural stands
A commonly noted variation in wood qualities within a species is that related to geographic source. Studies of wood of Pinus elliottii from natural stands have shown a definite trend toward decreased specific gravity from south to north and from east to west (Goddard and Strickland, 1962; Larson, 1957; Perry and Wang, 1958; Wheeler and Mitchell, 1959). Similar trends for specific gravity in Pinus taeda have also been found in the southeastern part of the United States by Mitchell and Wheeler (1959), Zobel and McElwee (1958) and Zobel et al. (1960). The last-mentioned study showed no trend for cellulose yields but reported a strong north to south trend in tracheid lengths, with trees from the North Coastal Plain having considerably shorter tracheids than those from the South Coastal Plain. In contrast to the significant geographic trends encountered in P. elliottii and P. taeda, P. serotina showed no such trends for either specific gravity or tracheid length in the much more limited geographic area of eastern North Carolina (McElwee and Zobel, 1963).
Geographic trends have also been reported for species other than southern pines. As early as 1916 Lee and Smith stated that tracheid length was greater from coastal Douglas fir (Pseudotsuga taxifolia) than from mountain Douglas fir. Myer (1930) found differences in wood from different geographic sources within the natural range of Pinus strobus and Tsuga canadensis, while Holst, in 1960, reported ecotypic as well as random variation in wood specific gravity of Picea glauca.
In a general paper on forest tree breeding, Hagman (1956) indicates that great differences have been established between different provenances and races in volume-weight of wood, its structure, cell size and other qualities. In working with Pseudotsuga taxifolia, Pinus sylvestris, Fagus sylvatica and Populus sp., Gohre (1958) found geographic variability but emphasized that there was so much variation in wood of individuals within a geographic area that it tended to obscure the between-geographic area differences.
A limited number of geographic studies have been made on wood properties of broadleaved species, a few of which will be cited here. In New York State, Valentine (1962) sampled 26 stands of Populus tremuloides for specific gravity and found real differences between stands. Thorbjornsen (1961) reported that fiber length differences in Liriodendron tulipifera were related to the site where the parent trees were growing. There are undoubtedly other studies of broadleaved species dealing with geographic variation but remarkably few of them have been published.
In summary, geographic variation in natural stands does not indicate anything about inheritance per se but it shows the extent and kind of variation patterns. To one who knows his species well, such information on natural stands will serve as a guide to the possible importance of genetics in determining wood properties and will result in more intelligently designed wood genetic research projects. To be of maximum efficiency, a study of natural variation should employ a nested sampling design so that geographic, site, tree-to-tree and within-tree variations are all determined at the same time while their relative importance is being assessed.
Geographic variation - provenance tests
It is interesting to note that most of the studies of wood in native stands have been made in North America where forest genetics is new and where large acreages of relatively undisturbed native species are available. Conversely, most provenance tests have been made in Europe, Australia, South Africa and other areas where exotic species or non-native races have been established. Wood characters of different provenances are of special interest because of the potential for rapid improvement.
There has been some confusion on the part of lay foresters about the role of genetics in studies where seed has been obtained from different sites or geographic areas and planted under different environments.
They tacitly assume that the wood properties of the provenance will be the same in its new environment as it is in its native habitat. This is often not true, and seed obtained from a natural stand with high specific gravity may produce seedlings with low specific gravity when grown under different environmental conditions. Such a reversal has often been cited as proof that genetic control of wood does not exist - an obviously false conclusion because for every wood character there is a genotype-environment interaction, which may vary in different environments.
Although many general provenance tests are under way, studies emphasizing wood properties of trees from seed of different geographic areas are still rather limited. Fortunately, findings are being reported at an increasing rate and soon considerable knowledge about wood of trees from different geographic areas will have been accumulated. In an early report, Klein (1957) found that Picea abies of German origin had higher specific gravity than that of Norwegian origin when grown in Scandinavia. Results to the contrary were found by Dietrichson (1961), who reported that southern provenances of both Pinus sylvestris and Picea abies had less summerwood than the native trees when grown in Norway. In contradistinction to both Klein's and Dietrichson's results, Knudsen (1956) failed to find differences among 8 provenances whose sources ranged from Norway to the U.S.S.R. Langlet (1938), in his summary of provenance tests of P. sylvestris, found that the northern provenances produced higher specific gravity than those from the south. In a comprehensive analysis of the international provenance tests near Nancy, France, significant differences were found in wood specific gravity among the 12 provenances of Picea abies tested by Parrot (1960), who concluded that genotypic differences in wood quality exist among certain populations.
For the exotic species Pinus pinaster grown in Australia, Bisset et al. (1951) found tracheid length differences which seemed to be tied to the seed source. Similarly, studies by Schütt (1958) of several provenances of Pinus contorta grown in Bavaria indicated certain cellulose and lignin differences related to provenance; coastal sources and those from good sites usually yielded the most cellulose and least lignin. In a recent study, Schüt (1962) found differences in both tracheid length and specific gravity among three sources of Pinus contorta.
An analysis of 4 provenances of Picea sitchensis (ranging from northern California to British Columbia) showed that 3 had similar specific gravity, but the fourth was significantly higher than the rest (Jeffers, 1959). In the statistical model of variation, the timber of the California source showed a consistently different pattern to that from other provenances. For Pinus ponderosa, Harris and Kripas (1959) found the provenance from California had lower specific gravity than that from British Columbia, when grown in New Zealand.
In the southeastern United States, Strickland (1960) found considerable differences in tracheid length and specific gravity among provenances of 22-year-old Pinus taeda grown in Georgia. The relationship between specific gravity of the progenies and latitude and longitude of origin of seed was positive and strong. In Tennessee, Thorbjornsen (1960) found differences in wood of P. taeda from different sources, but they did not follow the same pattern as in the natural stands. Conversely, in a 22-year-old provenance test of Pinus elliottii in Louisiana, considered by some foresters to be an especially "uniform" species, Derr and Enghardt (1960) found no significant wood specific gravity or tracheid length differences among sources. On another reportedly uniform pine species (Pinus resinosa), Rees and Brown (1954) found only one of 19 sources that possessed wood with a significantly different specific gravity. Haigh (1961) reported that provenance was the factor exerting the dominating influence on specific gravity of Pseudotsuga taxifolia, while for 18-year-old Picea glauca, Holst (1958) uncovered real differences among several provenances. In his study of the effect of provenance in Picea sitchensis, Dinwoodie (1963) obtained good evidence of racial variation, the California provenance having significantly different tracheid lengths from the others tested. In clones of Populus trichocarpa, Cech et al. (1960) found that differences in specific gravity and tracheid length were related to origin of seed.
This survey makes obvious the conclusion that for most species of forest trees differences in the wood are associated with racial differences or provenance. While several species showed only minor differences, usually the species occurring over a wide geographic range involving a diversity of habitats and environments not only have different wood properties in their native habitats but also show differences in wood when grown together under new and uniform environments. However, the characters found in the new environment may be quite different from those in the native habitat.
Inheritance of wood properties - individual trees
The prime interest in wood genetic studies centers on determining how wood characters are passed from parent to offspring, whether through vegetative or sexual means. For convenience, inheritance through sexual reproduction (often expressed as narrow-sense heritabilities) will be discussed separately from that through vegetative means.
Inheritance through sexual reproduction - specific gravity. It is possible to obtain an estimate of genetic variation either through control pollination (two parents) or open pollination (one parent) tests. The latter are less precise and give information only on the additive component of variance. The inheritances reported below will be shown as narrow-sense heritabilities, or as generalized observations in which the characters of the progeny are compared with those of the parents.
A number of studies report inheritance only in general terms and can be considered mainly as interesting observations serving as a guide. For example, Zobel and Rhodes (1957) studied progenies of several 12-year-old selfed trees, as well as open-pollinated progenies of Pinus taeda In every instance, high specific gravities of the progenies were associated with high parental specific gravities. In Texas. Brown and Klein (1961) found that when high specific gravity Pinus taeda parents were crossed, the 2-year-old progeny had a specific gravity of 0.39, while the progeny of high x low specific gravity parents averaged 0.35. They concluded that progeny groups did inherit specific gravity differences from the parent trees. In an unpublished report (1961), Klein and Brown showed results obtained from 17 control-pollinated crosses made among 8 trees selected for high and low specific gravity. Using 2-year-old seedling material, a regression of progeny specific gravities on mid-parent specific gravities was significant at the 1 percent level. About one fifth of the variability between parental combinations appeared to be genetically transmitted.
For Pinus radiata in Australia, Fielding and Brown (1960) determined variance components for a 6-year-old, open-pollinated progeny test. They obtained a heritability estimate of 0.2 which had a very large standard error. In a somewhat more comprehensive study, Squillace et al. (1962) estimated heritabilities for specific gravity and summerwood percent of both 14-year-old control- and open-pollinated progenies of Pinus elliottii, obtaining a narrow-sense heritability of 0.21 from the open-pollinated and 0.56 from control-pollinated progeny. This latter figure would seem unusually high for narrow-sense heritabilities of a character as complex as specific gravity, yet is similar to that found by Stonecypher et al. (1963) who obtained a heritability of 0.55 from progenies of 100 open-pollinated mother trees of Pinus taeda. The inheritance pattern in 3-year-old material was somewhat higher. Another recent report of Pinus taeda from Texas shows heritabilities of 0.37 and 0.49 for 2-year-old control-pollinated progeny (van Buijtenen, 1962), comparable to the high values obtained by Squillace and Stonecypher. Observing 17 6-year-old open-pollinated progeny groups, van Buijtenen states, "Differences were highly correlated with wood density of the female parent." Although he did not estimate heritability of the open pollinated progenies directly, he postulated that:" Its value is probably between 0.64 and 1.00." Progress from one generation of selection for specific gravity was estimated to be about 4 percent, based on a selection differential of one standard deviation, and assuming a heritability of 0.5.
TABLE 7. - HERITABILITIES FOR NINETEEN WOOD AND TRACHEID CHARACTERS IN FIVE-YEAR-OLD LOUISIANA AND GEORGIA Pinus taeda 1
|
Characters |
Louisiana material |
Georgia material |
|
|
Specific gravity of the core |
0.76 |
0.87 |
|
|
Percent summerwood in the core |
0.86 |
0.92 |
|
|
1959 summerwood |
|||
|
|
Percent in ring |
0.25 |
0.81 |
|
|
Width |
0.28 |
0.88 |
|
|
Specific gravity |
0.79 |
0.50 |
|
1959 summerwood tracheids |
|||
|
|
Double-wall thickness |
0.84 |
NC² |
|
|
Radial lumen diameter |
0.44 |
0.31 |
|
|
Radial width |
0.81 |
NC |
|
|
Tangential width |
0.59 |
0.27 |
|
|
Length |
0.97 |
0.85 |
|
|
Ratio of double-wall thick ness to radial width |
NC |
0.72 |
|
1959 springwood |
|||
|
|
Width |
NC |
0.74 |
|
|
Specific gravity |
0.04 |
0.35 |
|
1959 springwood tracheids |
|||
|
|
Double-wall thickness |
0.13 |
NC |
|
|
Radial lumen diameter |
0.59 |
0.68 |
|
|
Radial width |
0.76 |
0.65 |
|
|
Tangential width |
0.80 |
0.49 |
|
|
Length |
0.54 |
0.77 |
|
|
Ratio of double-wall thick ness to radial width |
0.15 |
0.67 |
1 From Goggans (1962)
² Not calculated because s 2 was estimated to be zero or minus.
TABLE 8. - SOURCES ON INHERITANCE OF SPECIFIC GRAVITY IN Pinus
|
Description of progeny |
Inheritance reported |
Literature source |
|||
|
Species |
Age |
Type of progeny 1 |
Broad sense |
Narrow sense |
|
|
Pinus radiata |
|
Clonal |
Strong |
- |
Fielding (1953) |
|
P. taeda |
12 |
Self and open poll. |
- |
Fairly strong |
Zobel and Rhodes (1957) |
|
P. radiata |
6 |
Open poll. |
- |
0.20 |
Fielding and Brown (1960) |
|
19 |
Clonal |
0.7 |
- |
Fielding and Brown (1960) |
|
|
20 |
Clonal |
0.5 |
- |
Fielding and Brown (1960) |
|
|
13 |
Clonal |
0.7 |
- |
Fielding and Brown (1960) |
|
|
P. taeda |
3 |
Open poll. |
- |
Fairly strong |
Smith (1962) |
|
P. sylvestris |
8-15 |
Clonal |
Strong |
- |
Ericson (1960) |
|
P. radiata |
8 |
Clonal |
0.45 to 0.75 |
- |
Dadswell and Wardrop (1960) |
|
P. taeda |
2 |
Control poll. |
- |
Fairly strong |
Brown and Klein (1961) |
|
P. taeda |
2 |
Control poll. |
- |
0.20 |
Klein and Brown (1961) |
|
P. radiata |
8 |
Clonal |
0.54 to 0.75 |
- |
Dadswell et al. (1961) |
|
P. elliottii
|
14 |
Open poll. |
- |
0.21 |
Squillace et al. (1962) |
|
14 |
Control poll. |
- |
0.56 |
Squillace et al. (1962) |
|
|
14 |
Clonal |
0.73 |
- |
Squillace et al. (1962) |
|
|
P. taeda |
2 |
Control poll. |
- |
0.37 to 0.49 |
van Buijtenen (1962) |
|
P. taeda |
6 |
Open poll. |
- |
0.64 to 1.0 (est.) |
van Buijtenen (1962) |
|
P. taeda
|
5 |
Louisiana (Open poll.) |
- |
0.76 |
Goggans (1962) |
|
5 |
Georgia (Open poll.) |
- |
0.87 |
Goggans (1962) |
|
|
(summerwood) |
5 |
Louisiana (Open poll.) |
- |
0.79 |
Goggans (1962) |
|
(summerwood) |
5 |
Georgia (Open poll.) |
- |
0.50 |
Goggans (1962) |
|
(springwood) |
5 |
Louisiana (Open poll) |
- |
0.04 |
Goggans (1962) |
|
(springwood) |
5 |
Georgia (Open poll) |
- |
0.35 |
Goggans (1962) |
|
P. elliottii |
5 |
Clonal |
0.46 to 0.73 |
- |
Zobel et al. (1963) |
|
P. taeda |
2 |
Open poll. |
- |
0.55 |
Stonecypher et al. (1963) |
1 Kind of pollination or vegetative propagation.
Van Buijtenen's data illustrate several important points. His high estimates for the open-pollinated progenies are not for a general population but for progenies of parents selected especially for high and low specific gravity. Theoretically, open-pollinated progenies should be half-sibs, but the stand histories and breeding behavior in Pinus taeda lead one to believe that there may be closer relationships among the progenies, and full-sibs and selfs may be present; also, there may be some relationship among the parents of the open-pollinated progenies. Van Buijtenen's inheritance values, therefore, are in a sense erroneous, but they must be considered as reflecting realizable heritability, that is, the inheritance in 6-year-old trees obtained by selecting from populations of Pinus taeda of the usual type from which selections will be made.
An important recent contribution dealing with inheritance of a number of wood properties is by Goggans (1962), who determined variance components for specific gravity of springwood and summerwood separately. He also studied percentage summerwood, length, width, wall thickness and various ratios of springwood and summerwood tracheids for 5-year-old open-pollinated Pinus taeda both from Louisiana and Georgia; the results are shown in Table 7. It is interesting to note the relatively strong genetic control of tracheid length, tracheid width, and specific gravity (except springwood specific gravity). Goggans reported heritabilities but cautioned that the actual numerical values may not be dependable because of the few progenies available for his studies. Relative rankings of characters are, however, of considerable interest.
In summary, heritabilities of specific gravity in the narrow sense based on young conifers are larger than expected for a quantitative and complex "character" dike specific gravity. These are summarized in Table 8. It has, however, been the experience of plant breeders that characters which have not been subject to strong selection pressures in the past have a high proportion of additive variance and respond well to selection. Specific gravity seems to follow this pattern; furthermore, it appears that heritabilities of specific gravity become stronger as older progenies become available for study.
Results from young progenies of Pinus taeda make it appear that with moderate to intense selection it should be possible to obtain an increase in specific gravity of 2 to 6 percent, based on overall specific gravity, while on the basis of the existing variation, it will be possible to gain from 8 to 15 percent of the possible improvement. Although too much emphasis should not be placed on the absolute values shown in Tables 7 and 8, their relative magnitudes give hope for good gains through selection.
Inheritance through asexual means - specific gravity. Many of the early studies dealing with specific gravity were made on grafts or rooted cuttings. For broadleaved species the most extensive studies have been done on Populus species where vegetative propagation is standard commercial practice. For such species, the broad-sense heritabilities obtained can be directly used to calculate gain in a breeding program. For species commonly propagated by seed, broad-sense inheritance values are useful as a research tool and an aid in determining the patterns that exist within a species indicating the desired direction and intensity of a breeding program, but cannot be used directly to predict gain.
Because natural regeneration can occur through clones in poplars, it is possible to get an estimate of broad-sense heritabilities from studies of wild populations. One of the most intensive of such studies was by van Buijtenen et al. (1959), who reported broad-sense heritability for specific gravity to be only about 0.17. The calculated gain in Populus tremuloides was a mere 2 percent, if the best 5 percent of the clones were selected; these results indicate that rewards from selecting poplar clones for specific gravity will be small. In a follow-up study, van Buijtenen (1962) found a range of variance ratios (an approximate estimate of gross heritability) from 0.17 to 0.43. He stated: "...it [is] likely that considerable improvement of the pulp and papermaking properties might be achieved by breeding methods," even though he shows specific gravity to have the least hope for improvement of the wood characters studied. For a similar study of triploid aspen, Einspahr et al. (1963) reports broad-sense heritabilities of specific gravity to be 0.38, a figure still rather low compared with conifers. Despite the low value for specific gravity per se, Einspahr feels that yield is under strong genetic control.
Definitive studies on inheritance of wood properties in other broadleaved species are disappointingly few, although a number of papers deal with general inheritance patterns. For example, Cech et al. (1960) demon: strafed statistical differences in the wood of clones of Black cottonwood (Populus trichocarpa) with a large clonal influence on specific gravity. Kennedy and Smith (1959), working on both Populus trichocarpa and the hybrid Populus "Regenerata," found real differences among 1-year-old trees, and suggest that improvement can be achieved by selecting clones of high specific gravity. In still another study of Populus trichocarpa, Gabriel (1956) found quite large statistically significant differences between clones. He concluded that much of the observed clonal differences may be genetically controlled.
Inheritance of wood properties in vegetatively propagated coniferous material has received the most study of all. Broad-sense heritabilities were summarized by Zobel (1961), and only brief references will be made to the earlier studies here. First indications of intensity of broad-sense heritabilities, were reported by Fielding (1953), who found that clone 22 of Pinus radiata consistently produced wood approximately 20 percent heavier than clone 41. Working on Pinus radiata, Fielding and Brown (1960), and Dadswell et al. (1961) reported broad-sense heritabilities from 0.45 to 0.75 for trees of different ages (6 to 20 years), representing various positions within the tree, and representing either "whole wood" or summerwood only. In a recent paper, van Buijtenen (1962) reported broad-sense heritabilities, of 0.17 for 1-year-old Pinus taeda wood and 0.64 for 5-year-old wood of grafts. This latter figure is quite close to that obtained by Zobel et al. (1963) for 5-year Pinus elliottii, in which broad-sense heritabilities for 258 grafts representing 38 clones ranged from 0.46 to 0.73, depending on location of tests and whether or not data were pooled over the two locations where the grafts were grown. For 12- to 14-year-old Pinus elliottii, Squillace et al. (1962) obtained a broad-sense heritability of 0.73 for specific gravity.
A number of authors have determined inheritance patterns for clonal material without converting their results to broad-sense heritabilities. One of the most comprehensive of these was by Ericson (1960, 1961) on the relation between wood of grafts and wood of the parent trees from which they were obtained. He found strong, positive correlations for Pinus sylvestris on 533 grafts (72 clones in 3 clone collections) and similar relationship between original tree and graft tree for Picea abies. He emphasized the importance of selecting for wood specific gravity in practical tree breeding.
In summary, the determination of broad-sense heritabilities has stimulated the interest and enthusiasm for research in the field of wood genetics, even though the results cannot always be used directly to determine gain. The report of clonal differences by Fielding (1953) lent encouragement to men interested in wood genetics and actually "triggered" a number of later studies. Although broad-sense heritabilities indicate the maximum genetic gain, rarely realizable by regeneration through seed, their magnitude, with the possible exception of poplars, has shown the potential for reasonably high narrow-sense values. The increasing trend with age has been interesting, and a narrowing of the difference between broad- and narrow-sense heritabilities for older conifers provides the basis for some intriguing speculations.
Inheritance of wood properties, other than specific gravity, through sexual means. Although most of the discussion so far has dealt with specific gravity, there is considerable information available regarding inheritance of other wood properties. The most widespread reports on conifers have dealt with tracheid length; this character has shown a consistently high pattern of inheritance. In one of the earlier papers, Echols (1955) reported tracheid length of Pinus elliottii to be under rigid genetic control, the inheritance pattern indicating that tracheid length is governed by a multiple gene series. Echols showed a direct linear relation between fibrillar angle and tracheid length; by inference, therefore, fibrillar angle should be under rigid genetic control. Jackson and Greene (1957) also reported that Pinus elliottii parents with long tracheids produced progeny with longer tracheids than did parents with short tracheids. In most instances, tracheid lengths of progeny were intermediate to that of the parents, but in some instances length seemed to be influenced more by the female than the male parent. There have been many other reports of general control of tracheid lengths. For example, Kramer (1957) and Schreiner (1958) both postulate strong genetic control of tracheid length, based on variation studies.
An early specific instance of inheritance of tracheid length was reported for the larch hybrid Larix x eurolepis by Chowdhury (1931), who found springwood tracheids resembling those of the female parent while summerwood tracheids were intermediate between the parental species. He also found tangential diameter of tracheids and thickness of tracheid walls to be intermediate between the parents. Another study reporting inheritance values for tracheid characters was that by Goggans (1962). He obtained narrow-sense heritabilities for tracheid length of springwood of 0.54 and of summerwood of 0.57 in Louisiana, U.S.A. Values obtained for wall thickness and tracheid width of both springwood and summerwood of 5-year-old Pinus taeda in general revealed inheritance patterns much stronger than expected (see Table 7). These patterns raise hopes for achieving real improvement, particularly in length-width ratios which are considered so important in paper making.
Another wood quality for which many studies have shown possibilities for genetic improvement is spiral grain. There are numerous papers dealing with this subject, but those published for Pinus roxburghii by Champion (1945) and coworkers are perhaps most appropriate in pointing out the possible genetic control of spiral grain. Self-pollinated, twisted P. roxburghii parents produced seedlings that were 68 to 82 percent twisted, observable in the cotyledon stage, leading Kadambi and Dabral (1955) to conclude that twist is a dominant character. Schmucker (1956) concluded that spiral grain in Fagus sylvatica is the rule and, therefore, probably results from inherent factors; he feels that it is a dominant genetic character. A similar conclusion was reached by Northcott (1957) who studied 6 broadleaved and 6 conifer species. The variation and complexity of spiral grain and its change with age have been well summarized by Noskowiak (1960).
The concept of strong inheritance of spiral grain is challenged by Paul (1953), who feels that the spiral is caused by some factor other than heredity. In his opinion, interlocked grain is more likely to be strongly controlled genetically than spiral grain. Webb (1963) has observed pronounced differences in interlocked grain among trees of Liquidambar styraciflua, but he has not yet been able to assess directly the intensity of inheritance of this character.
Varied grain patterns have been subjected to considerable study. In Finland, Ahola (1952) investigated 311 gnarled trees of Pinus sylvestris and 23 gnarled trees of Picea abies, concluding that there is some evidence that this pattern is inherited. Open-pollinated seed of curly-grained parents of Betula verrucosa were found by Heikinheimo (1951) to produce about 50 percent curly-grained progeny, only a few of which grow to tree size, the rest remaining shrubs. Johnsson (1950) in Sweden concluded that wavy grain is conditioned by an hereditary disturbance in the function of the cambium, control crosses showing a high frequency (but less than 100 percent) of curly birch. Sometimes open-pollinated trees produced a high proportion of curly progeny. Based on his own research and that of others, Ruden (1954) proposed the varietal name Betula verrucosa var. maserica for curly birch produced by genetic peculiarities in bark growth.
There are scattered reports on inheritance of other morphological wood characters but very few indicate more than a tendency toward inheritance, usually toward intermediacy between parents. Typical is the study by Pryor et al. (1956), who reported that wood properties of most Eucalyptus hybrids were intermediate or the same as the parents, and seemed to be under multiple factor control.
The chemical character of wood is one of the most important with which the forest geneticist must deal. Studies of variation in this important character to date have dealt mainly with broad-sense heritabilities, few results being available on sexual inheritance. Inheritance of the resin content of wood has not been reported, although numerous studies report strong inheritance for resin yield by tapping. Whether this applies to the resin contained in the wood is not known, but it seems reasonable that the tendency to strongly resinous normal wood of trees might be quite strongly controlled genetically.
Inheritance of wood properties, other than specific gravity, through asexual means. Broad-sense heritabilities of tracheid lengths and other tracheid and chemical characters have been shown by a number of authors, including Dadswell and Wardrop (1960) and Dadswell et al. (1961), for 8-year clones of Pinus radiata. They reported the broad-sense heritabilities for tracheid length to be 0.73, for spiral grain 0.66, and cellulose yield 0.29. In their recent study on Pinus elliottii, Zobel et al. (1963) found heritabilities in tracheid lengths of 0.56 for 5-year grafts. No relationship between the wood of the scion and the root stock or tree to which it was grafted was found. Greene and Carmon (1962) found differences in tracheid lengths between clonal lines in Pinus echinata.
Strong inheritance of fiber length was revealed in studies of Populus tremuloides by van Buijtenen (1959) and van Buijtenen et al. (1962). Fiber length in triploid aspen had broad-sense heritabilities of 0.86 at age 30, according to Einspahr et al. (1963), who also calculated inheritance of handsheet properties, obtaining values as high as 0.90 for tear factor, 0.83 for bursting strength and 0.57 for tensile strength. They concluded that there must be strong genetic control over fundamental fiber properties which, in turn, influence handsheet strength properties. In Populus Schönbach (1960) found real differences between clones for cellulose yield and suggested that cellulose content should be taken into account in selective breeding.
Much has been done with figured wood by asexual propagation. Trials with 8 figured trees resulted in curly progeny of Acer rubrum, Liriodendron tulipifera and Juglans nigra, according to Bailey (1948). In 1953, Tellerup was able to prove distinct individual differences in the shape of wood rays in Fagus sylvatica from several different clones, and Walters (1951) found the figured character of Juglans nigra was transmitted by grafting. In certain hardwoods (Quercus alba, for example) Zabel (1956) found more decay in the test blocks of decay susceptible trees than from decay resistant trees. He observed this result also in Robinia pseudoacacia and quotes other authors as stating that durability is a "rigidly controlled" wood character.
Breeding trees for better wood involves both the use of genetics and the manipulation of the environment. As indicated earlier, a complete treatment of improvement through both environmental and genetical means cannot be accomplished in a summary chapter. In dealing principally with the inheritance phase, the incompleteness of this treatment must be recognized and it must be considered that one of the best ways to improve wood is simply to improve tree habit and growth. The potentials for this are very great. For example, improvement of the inherent straightness of trees automatically improves the quality of the wood for any use to which it will be put. There are a host of papers supporting this as well as the high degree of inheritance of stem straightness. Space does not permit citations of the papers, but they agree universally that straightness is one of the most strongly controlled genetic characters, making possible large and rapid improvements in stem characters through breeding. Other papers emphasize the importance of straightness for wood qualities. It is quite possible that for some species the greatest improvement in wood quality will be achieved by breeding for better stem characters.
Exactly the same reasoning applies for branch characters such as size, branch angle, rapidity or ease of pruning, sprouting ability, and so on. The extent and kinds of improvement possible through breeding trees which have better crowns, smaller branches or better branch angles needs critical evaluation. Similar assessment is needed for growth rate and general growth pattern of the tree because, for certain-species and certain products, much depends on controlling growth characters which can be affected by genetic manipulation.
Is breeding for wood qualities per se worthwhile?
Any method or system of breeding is expensive and should not be undertaken, at least on a large-scale, unless the results will justify the effort and cost. As tree breeders, the fact must be faced that, although wood is the final product, in many instances the development of faster growth, better growth habit, better adaptability, and better resistance to pests, will continue to be major objectives if the improved trees contain wood as good as at present. A question of the first importance is to determine whether or no breeding for better growth habit and resistance characters has caused a deterioration in wood quality which will make the improved trees less desirable. Because there is this tendency under certain conditions, especially when moving provenances, the breeder will have attained his objective simply by preserving present wood qualities and preventing degrade in the improved tree.
The second question, except in certain special cases, is to determine whether wood qualities can be improved within the framework of better habit, faster growth, greater hardiness, etc. For those species that must be reproduced by seed it may be impracticable to breed for wood qualities first and then to attempt to improve the growth and form of the trees with the improved wood. It seems much more logical to improve wood within the framework of good growth and habit. The most encouraging part of the wood breeding picture is that, for nearly every wood character studied, there is a huge amount of variation even within groups of trees rigidly selected for growth and form. Studies in North Carolina and is many other places indicate that the range in wood qualities is nearly the same among trees selected for growth and form as for unselected populations. This aspect is most heartening because, if one selects for wood within the framework of growth and form, there are many fewer trees from which selections can be made.
Another encouraging point is that the major share of the variance in wood quality appears to be additive in nature, thus reacting positively to intensive selection and sexual propagation. As with all aspects of breeding, the problems are much simpler when vegetative propagation is involved. The major danger arises from having too few clones of the desired types, since this paucity may open the door to catastrophes that sometimes have fallen on large areas planted to a sin ale or few genotypes. The heritabilities that have been obtained to date in both the broad and the narrow sense, are satisfyingly large. Variation within the populations is wide, yielding a large selection differential. For certain characters, especially specific gravity, heritabilities increase with age of progeny, at least during the early years.
Even with all these favorable factors the tree breeder should not be deluded into hoping for, or predicting, large gains in one generation of selection. He must ask: What value is an increase in tracheid length of 0.5 mm? What value is an increase in dry weight per cord of 45 kilograms (100 pounds)? What are 10 percent more trees in the stand with desired wood qualities worth? Such gains can be achieved when one is working with the narrow- and broad-sense heritabilities and selection differentials reported in this paper. Another question is: What value is greater uniformity? The genetic approach can result in an improvement rarely stressed - improvement of uniformity. Complete uniformity will not be obtained; the inherent growth patterns within trees and the effect of environmental differences preclude this. But it is possible to produce much more uniform wood through breeding, especially in those species reproduced vegetatively. It is safe to say that improved uniformity alone, whether or not accompanied by any other improvement, will pay the cost of breeding for wood qualities many times over.
Breeding for growth habit, growth rate and pest resistance will continue to receive much attention in tree improvement programs. But, with the pattern of inheritance and variation known for wood qualities, it is hard to see how any forest tree breeder who hopes to make the greatest improvement can afford not to include breeding for wood quality. Improved wood quality, more uniform wood, or in some instances just a maintenance of the status quo will repay the effort many times over. Percentage improvements may seem small; yet when one considers that all the wood harvested every year for trees that have occupied the site for many years will have this improvement, it becomes impressive indeed. One co-operator in the Industry-North Carolina State Co-operative Tree Improvement Program has estimated that an improvement in yield of just 1 percent would amount to more than U.S. $1 million a year for one pulp mill.
Evaluating the environment
The foregoing discussion shows clearly that certain wood characters are inherited and that a genetic gain in wood quality can be realized by a judicious breeding program. It is equally apparent, however, that the local environment strongly modifies wood quality by its influence on the overall growth pattern of the tree. A specific environment will bring forth a fairly definite outward expression of the basic genetic constitution of the tree in terms of wood quality. But the genetic potential of a tree permits a rather wide range of phenotypic expression, and every minute change in the environment will result in a different growth response and, consequently, a slightly altered wood quality. Thus, a particular genotype may exhibit an array of wood quality assessments when grown under a range of environments. When one considers the multiplicity of genotypes and environments that exist in nature within a species range, the potential diversity in wood quality can be readily appreciated. As pointed out previously, however, this diversity is most encouraging for the tree breeder, as it provides real opportunities for selection of desirable wood quality attributes.
Although genotype and environment are inextricably united, in the sense that no tree can be grown completely devoid of its environment, it is possible to evaluate genotype "independent" of environment by suitable experimental designs and statistical controls. This approach forms the basis for the heritability estimates previously described. A somewhat similar approach has been used to isolate the influences of environment, although seldom has the genotype been rigidly controlled. Nevertheless, if one assumes a fairly uniform genetic response because of the generally restricted units of sampling, one may accept as valid the results of the many excellent studies relating environment to wood quality. Numerous experiments of this nature have been conducted in past years, and considerable information has accumulated indicating a pronounced modifying influence of environment on tree growth and wood properties. No attempt will be made at this time to make a comprehensive review of this vast literature. The subject has been reviewed from numerous points of view over the years. For a more detailed discussion of the subject and a listing of pertinent literature citations, the reader is referred to the following papers: Lassila (1930); Klein (1933); Spurr and Hsiung (1954); Trendelenburg and Mayer-Wegelin (1955); Larson (1957); Knigge (1958); Zobel et al. (1960); Dinwoodie (1961); Goggans (1961); Tappi (1962).
Among the most widely recognized environmental factors contributing to variability in wood quality are weather conditions, stand structure, site quality, geographic locality, and the amelioration of stand and site conditions by silviculture. The influence of several of these environmental factors cannot be readily segregated under natural field conditions. For example, natural stands identical in spacing and structure are impossible to locate, and comparable site qualities can only be attained with certainty on closely adjacent areas. The assessment of wood quality is further complicated by the vagaries of changing weather conditions during a growing season, since these effects are superimposed on all the other environmental factors. Interactions between prevailing weather and site are particularly troublesome when the sites are separated-in distance; this applies not only to natural stands but also to seed source and clonal plantations. These remarks are not intended to discourage research activity in this vital area of environmental influences, but rather to emphasize a few of the precautionary measures required for such studies.
In spite of the many complicating factors briefly cited above, the influence of environment on tree growth is so strong that very striking qualitative differences in wood quality can be demonstrated. Meaningful quantitative data can also be obtained in some cases, particularly where carefully collected measurement data are available on climatic, stand, or site variables.
The determination of heritabilities and the assessment of genetic components of wood quality require that the contributions of environment be fully recognized and evaluated. This may be accomplished either by statistical procedures that allow for environmental variation or by experimental designs that hold environmental variation to a minimum.
In considering wood quality as influenced by both inheritance and environment, a dynamic relationship is being studied. As the environment changes, wood quality changes, and underlying all of these alterations in tree growth is an inherited potential that predisposes each tree to respond in its own characteristic manner.
AHOLA, V. K. 1952. On gnarly trees. Commun. Inst. for Fenn., 40 (18): 1-10.
BAILEY, L. F. 1948. Figured wood: a study of methods of production. J. For., 46 (2): 119-125.
BISSET, I. J., DADSWELL, H. E. & WARDROP, A. B. 1951. Factors influencing tracheid length in conifer stems. Aust. For., 15 (1): 17-30.
BROWN, C. L. & KLEIN, J. 1961, Observations on inheritance of wood specific gravity in seedling progeny of Loblolly pine. J. For., 59 (12): 898-899.
BUIJTENEN, J. P. VAN. 1962. Heritability estimates of wood density in Loblolly pines. Tappi, 45 (7): 602-605.
BUIJTENEN, J. P. VAN, EINSPAHR, D. W. & JORANSON, P. N. 1959. Natural variation in Populus tremuloides Michx. Tappi, 42, (10): 819-823.
BUIJTENEN, J. P. VAN, EINSPAHR, D. W. & PECKHAM, J. 1962. Natural variation in Populus tremuloides Michx. II. Variation in pulp and papermaking qualities. Tappi, 45: 58-61.
CECH, F. C. & ZOBEL, B. J. 1960. What is inherited - how can we tell? Forest Farmer, July. 2 p.
CECH M. Y., KENNEDY, R. W. & SMITH, J. H. G. 1960. Variation in some wood quality attributes of one-year-old black cottonwood. Tappi, 43 (10): 857-858.
CHAMPION, H. G. 1945. Genetics of forestry. Quart. J. For., 39 (2): 74-81.
CHOWDHURY, A. 1931. Anatomical studies of the wood of a hybrid larch. J. For., 29 (5): 797-805.
DADSWELL, H. E. & WARDROP, A. B. 1960. Some aspects of wood anatomy in relation to pulping quality and to tree breeding. J. Aust. Pulp and Paper Ind. Tech. Assoc., 13 (5): 161-173.
DADSWELL, H. E., FIELDING, J. M., NICHOLS, J. W. P. & BROWN, A. G. 1961. Tree-to-tree variations and the gross heritability of wood characteristics of Pinus radiata. Tappi, 44 (3): 174-179.
DERR, H. J. & ENGHARDT, H. 1960. Is geographic seed source of Slash pine important ? Sth. Lumberm., 201 (2513): 95-96.
DIETRICHSON, J. 1961. Using southern provenances in Norway. Norsk Skogbruk, 6: 229-231.
DINWOODIE, J. M. 1961. Tracheid and fibre length in timber: a review of the literature. Forestry, 34: 125-144.
DINWOODIE, J. M. 1963. Variation in tracheid length in Picea sitchensis Carr. Dept. of Sci. and Ind. Res. Special Report, 16: 155.
ECHOLS, R. M. 1955. Linear relation of fibrillar angle to tracheid length, and genetic control of tracheid length in slash pine. Trop. Woods, 102: 11-22.
EINSPAHR, D. W., VAN BUIJTENEN, J. P. & PECKHAM, J. R.1963. Natural variation and heritability in Triploid aspen. Tappi, 46.
ERICSON, B. 1960. Studier över den ärftliga volymviktsvariationen hos tall och gran [Studies of the genetical wood density variation in Scots pine and Norway spruce]. Forest Res. Inst. of Sweden, Dept. of Forestry, Yield Research, 4: 152.
ERICSON, B. 1961. Forest tree breeding with a view to raising the yield of pulp. Some preliminary results. Teknisk Vetenskaplig Forshning, 32 (4): 194-203.
FIELDING, J. M. 1953. Variations in Monterey pine. Bull. For. Timb. Burl Australia, No. 31. 43 p.
FIELDING, J. M. & BROWN, A. G. 1960. Variations in the density of the wood of Monterey pine from tree to tree. Forestry and Timber Burl, Commonw. Australia, Leaflet No. 77. 28 p.
GABRIEL, W. 1956. Preliminary report on clonal differences in the wood and phloem of Populus deltoides and P. trichocarpa. Proc. 3rd Northeast. For Impr. Conf., p. 33-35.
GODDARD, R. E. & STRICKLAND, R. K. 1962. Geographic variation in wood specific gravity of slash pine. Tappi, 45 (7): 606-608.
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. Tech. Rpt. 11, School of Forestry, N. C. State College, Raleigh. 56 p.
GOGGANS, J. F. 1962. The correlation, variation, and inheritance of wood properties in loblolly pine (Pinus taeda L.) Tech. Rept. 14, School of Forestry, N. C. State College, Raleigh. 165 p.
GOHRE, K. 1958. The distribution of specific gravity in the stem and its variation with growth region and tree site. Holz Roh- u. Verkstoff, 16 (3): 77-90.
GREENE, J. T. & CARMON, J. L. 1962. Variation of tracheid length in clonal lines of short-leaf pine. Georgia Forest Research Council, Georgia. For. Res. Paper 13. 6 p.
HAGMAN, M. 1956. Maddallisuuksista parantaa puun käyttöominaisuuksia rodunjalostustoiminnalla [On the possibilities of improving the utilization value of trees through forest tree breeding]. Paper and Timber, 38 (2): 53-56.
HAIGH, W. 1961. The effect of provenance and growth rate on specific gravity and summerwood percent- age of young Douglas fir. U.B.C. Forest Club, Research Committee, 19: 57.
HANSON, W. D. 1963. Heritability. Statistical genetics and plant breeding. Humphrey Press, p. 125-140.
HARRIS, J. M. & KRIPAS, S. 1959. The physical properties of two provenances of ponderosa pine grown in Kaingaroa State Forest. N.Z. For. Res. Notes, 16: 3-16.
HEIKINHEIMO, O. 1951. Kokemuksia visakoivun kasvatuksesta [Experiments in growing curly birch]. Commun. Inst. For. Fenn., 39 (5): 26.
HOLST, M. 1958. Thoughts on wood density. Proc. 6th. Meeting Committee For. Tree Breeding in Canada, Part II, S-31 and S-32.
HOLST, M. 1960. Forest tree breeding and genetics at the Petawawa Forest Experiment Station. Proc. 7th Meeting Committee For. Tree Breeding in Canada, Part II, K-1 - K-27.
JACKSON, L. W. R. & GREEN, J. T. 1957. Hereditary variations in slash pine tracheids. Proc. 4th South. Tree Impr. Conf., Univ. of Georgia, p. 23-26.
JEFFERS, J. N. R. 1959. Regression models of variation in specific gravity in four provenances of Sitka spruce. J. Inst. Wood Sci., 4: 4459.
JOHNSSON, H. 1950. Avkommor av masurbjörk [Offspring of curly birch]. Arsberätt Fören. Växtföräd. Skogsträd, p. 18-29.
KADAMBI, K. & DABRAL S. N. 1955. On twist in Chir (Pinus longifolia Roxb.). Indian For., 81 (1): 58-64.
KENNEDY, R. W. & SMITH, J. E. G. 1959. The effects of some genetic and environmental factors on wood quality in poplar. Res. Note No. 19, Faculty of Forestry, Univ. of British Columbia. 2 p.
KLEIN, G. G. 1933. Untersuchungen über die Qualität des Fichtenholzes [Studies of the quality of spruce wood]. Medd. fra det Norske Skagsførsaksvesen, 5: 197-348.
KLEIN, G. G. 1957. Kvalitetsundersøkelser av norsk og tysk gran [The quality of Norway spruce (Picea abies) of Norwegian and German origin]. Medd. fra det Norske Skogsforsøksvesen, Vollebekk, 16 (48): 290-314.
KLEIN, J. & BROWN, C. L. 1961. Preliminary observation on the inheritance of wood specific gravity in loblolly pine seedlings, p. 1-10 (unpublished).
KNIGGE, W. 1958. Untersuchungen über die Beziehunzen Zwischen Holzeigenschaften und Wuchs der Gastbaumart Douglasie (Pseudotsuga taxifolia Britt.). Schriftenreihe forstl. Fak. Univ. Göttingen 20. 101 p.
KNUDSEN, M. V. 1956. A comparative study of some technological properties of Norway spruce in a provenance test. Proc. 12th Congr. IUFRO, Oxford, p. 1-7.
KRAMER, P. R. 1957. Tracheid length variation in Loblolly pine. Texas For. Serv. Tech. Rept. No. 10. 22 p.
LANGLET, O. 1938. Proviensförsök med olika trädslag [Provenance tests with various wood species. A summary and discussion of results to date]. Särtryck ur Svenska Skogsvårdsföreningens Tidskrift, 12: 1-278.
LARSON, P. R. 1957. Effect of environment on the percentage of summerwood and specific gravity of slash pine. Yale Univ. Bull. No. 63. 87 p.
LASSILA, I. 1930. On the influence of forest type on weight of wood. Acta for. Fennica, 36: 1-125.
LEE, H. N. & SMITH, E. M. 1916. Douglas fir fiber with special reference to length. For. Quarterly, 14 (4): 671-695.
LUSH, J. L. 1949. Heritability of quantitative characters in farm animals. Proc. 8th Inter. Congr. Hereditas, Suppl. Vol., p. 356-375.
MCELWEE R. L. & ZOBEL, B. J. 1963. Some wood and growth characteristics of pond pine. Proc. SAF Tree Impr. Workshop, Macon, Gal, Oct. 1962, p. 18-25. Publication No. 22, Southern Forest Tree Impr. Committee.
MITCHELL, U. L. & WHEELER, P. R. 1959. Wood quality of Mississippi's pine resources. Rept. No. 2143. U. S. For. Prod. Lab., Madison. 11 p.
MYER J. E. 1930. The structure and strength of four N. A. woods as influenced by range, habitat and position in the tree. N. Y. State College For. Tech. Publ. 31. 39 p.
NORTHCOTT, P. L. 1957. Is spiral grain the normal growth pattern? For. Chron. 33, (4): 335-352.
NOSKOWIAK, A. F. 1960. Spiral grain patterns in red pine and relationship of age and radial growth rate to change of grain angle. Ph. D. thesis, State Univ. College of Forestry, Syracuse Univ. N.Y.
PARROT, L. 1960. De la variabilité génétique de la densité du bois chez l'épicéa (Picea excelsa Link.). Ann. Ec. Eaux For. Nancy, 17 (3): 267-334.
PAUL, B. 1953. If. Forest genetics in relation to wood quality. Proc. Lake States For. Gen. Conf., Misc. Rept. No. 22, p. 55-59.
PERRY, T. O. & WANG, CHI WU. 1958. Variation in the specific gravity of slash pinewood and its genetic and silvicultural implications. Tappi, 41 (4): 178-180.
PRYOR, L. D., CHATTAWAY, M. M. & KLOOT, N. H. 1956. The inheritance of wood and bark characters in Eucalyptus. Aust. J. Bot., 4 (3): 216-239.
REES, L. W. & BROWN, R. M. 1954. Wood density and seed source in young plantation red pine. J. For., 52 (9): 662-665.
RUDEN, T. 1954. Om Valdbjörk og endel andre unormale veddanalser hos björk (Brown curly birch and some other abnormal wood formations in birch). Medd. Norske Skogforsoksv., 12 (3): 451-505.
SCHMUCKER, T. 1956. Forstgenetik: einige Befunde am Rande. Forstwiss, Cbl., 75: 1-2.
SCHÖNBACH, H. 1960. Contributions to poplar research No. IV Results of further studies in cellulose content and specific gravity of gray poplar clones. Wissenschaftliche Abhandlungen, 44: 83-98.
SCHREINER, E. J. 1935. Possibilities of improving pulping characteristics of pulpwoods by controlled hybridization of forest trees. Paper Trade Jour., Tech. Sec. C: 105-109.
SCHREINER, E. J. 1958. Possibilities for genetic improvement in the utilisation potentials of forest trees. Silvae Genet., 7 (4): 122-128.
SCHÜTT, P. 1958. Variations in the cellulose and lignin content of some Pinus contorta strains grown in West Germany. Silvae Genet., 7 (2): 65-69.
SCHÜTT, P. 1962. Individuelle und bestandesweise Schwankungen der Holzdichte und der Faserlänge bei Pinus contorta Z. fur die Erzeugung von Holzstoff, Zellstoff, Papier und Pappe. Chemische Technologie der Cellulose, 16 (11): 671-676.
SMITH, D. E. 1962. An investigation into the specific gravity relationship between limb sections and the bole or increment core of young loblolly pine (Pinus taeda L.) trees. M. S. thesis, N.C. State College, Raleigh, N.C., 1-58 (unpublished).
SPURR, S. H. & HSIUNG, W. 1954. Growth rate and specific gravity in conifers. J. For., 52: 191 200.
SQUILLACE, A. E., ECHOLS, R. M. & DORMAN, K. W. 1962. Heritability of specific gravity and summerwood percent and relation to other factors in slash pine. Tappi, 45 (7): 599-601.
STERN, K. 1958. Kombinationseignung hinsichllich der Wachstums-Ergebnisse eines Modellversuches mit Antirrhinum majus L. Silvae Genet, 7 (2): 41-57.
STERN, K. 1960. Über einige populationsgenetische Probleme der Auslese bei Forstpflanzen. Given at a seminar.
STONECYPHER, R., CECH, F. & ZOBEL, B. J. 1963. Inheritance of specific gravity in two and three year old seedlings of loblolly pine. Tappi, 46.
STRICKLAND, R. K. 1960. Geographic variations in specific gravity and tracheid length of loblolly pine. M.S. thesis, Univ. of Georgia, Athens, U.S.A.
TAPPI FOREST BIOLOGY COMMITTEE. 1962. The influence of environment and genetics on pulpwood quality: an annotated bibliography. Tappi Monog. 24. 316 p.
TELLERUP, E. 1953. Individual differences in the shape of wood rays in Fagus sylvatica L. A wood anatomical investigation. Royal Vet. and Agric. Col. Yearbook, p. 147-157.
THORBJORNSEN, E. 1960. Variation in loblolly pine (Pinus taeda L.). Ph. D. thesis, N.C. State College, Raleigh, N.C.
THORBJORNSEN, E. 1961. Variation in density and fiber length in wood of yellow poplar. Tappi, 44 (3): 192-195.
TODA, R. 1961. Studies on the genetic variance in Cryptomeria. Bull. Gov. For. Expt. Sta., 132: 1-46.
TRENDELENBURG, R. & MAYER-WEGELIN, H. 1955. Das Holz als Rohstoff. Munich, Carl Hanser Verlag. 541 p.
VALENTINE, F. A. 1962. Natural variation in specific gravity in Populus tremuloides in northern-New York. Proc. 9th Northeastern For. Tree Impr. Conf., Syracuse, N.Y., p.17-24.
WALTERS, C. S. 1951. Figured walnut propagated by grafting. J. For., 49 (12): 917.
WEBB, C. A. 1963. Natural variation in specific gravity, fiber length and interlocking grain of the wood of sweetgum (Liquidambar styraciflua L.). Ph. D. thesis, School of Forestry, N.C. State College, Raleigh. N.C. (unpublished)
WHEELER, P. R. & MITCHELL, E. L. 1959. Specific gravity variation in Mississippi pines. Proc. 5th South. For. Tree Impr. Conf., p. 87-96.
ZABEL, R. 1956. Decay resistance variations within north eastern forest tree species. Proc. 3rd Northeast. For. Tree Impr. Conf., p. 13-17.
ZOBEL, B. J. 1956. Genetic growth and environmental factors affecting specific gravity of loblolly pine. (Comments by E. J. Schreiner). For. Prod. J., 6 (10) 442-447.
ZOBEL, B. J. 1961. Inheritance of wood properties in conifers. Silvae Genet., 10 (3): 65-70.
ZOBEL, B. J. & RHODES, R. R. 1957. Specific gravity indices for use in breeding loblolly pine. For. Sci., 3 (3): 281-285.
ZOBEL, B. J. & MCELWEE, R. L. 1958. Natural variation in wood specific gravity of loblolly pine, and an analysis of contributing factors. Tappi, 41 (4): 158-161.
ZOBEL, B. J., THORBJORNSEN, E. & HENSON, F. 1960. Geographic, site and individual tree variation in wood properties of loblolly pine. Silvae Genet., 9: 149-158.
ZOBEL, B. J., COKE, D. & STONECYPHER, R. 1963. Wood properties of clones of slash pine. Proc. SAF Tree Impr. Workshop, Macon, Ga. Oct. 1962, p. 32-39. Publication No. 22, Southern Forest Tree Impr. Committee.
