Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Intraspecific variation in the IUFRO 16 year old Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco) :… Kantarl, Mustafa 1989

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1989_A6_7 K36.pdf [ 5.98MB ]
Metadata
JSON: 831-1.0075400.json
JSON-LD: 831-1.0075400-ld.json
RDF/XML (Pretty): 831-1.0075400-rdf.xml
RDF/JSON: 831-1.0075400-rdf.json
Turtle: 831-1.0075400-turtle.txt
N-Triples: 831-1.0075400-rdf-ntriples.txt
Original Record: 831-1.0075400-source.json
Full Text
831-1.0075400-fulltext.txt
Citation
831-1.0075400.ris

Full Text

INTRASPECIFIC VARIATION IN THE IUFRO 16 YEAR OLD DOUGLAS-FIR (PSEUDOTSUGA MENZIESII (MIPS.) FRANCO) PROVENANCE-PROGENY TRIAL IN HANEY, BRITISH COLUMBIA by MUSTAFA KANTARLI B.Sc.F., Blacksea Technical University, 1981 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS OF THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES in THE DEPARTMENT OF FORESTRY We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA Apri l , 1989 © Mustafa Kantarli, 1989 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada DE-6 (2/88) ABSTRACT Variation in height, dbh and volume was analyzed after 16 years of th th growth. Results were compared with the 4 and 7 growing season measurements for height. This trial comprised 464 open-pollinated families representing 58 provenances planted at the University of British Columbia Research Forest in Haney, B.C. in 1971. Phenotypic variations between and within provenances, additive genetic variances, heritabilities, genetic gain and juvenile * mature correlations were estimated for the three growth variables according to zone groupings. The effects of provenance by block and family by block interactions were evaluated. Expected reductions from the genetic gain caused by the interactions were calculated. Relationships among growth variables and growth variables versus geographical variables were investigated by simple and multiple linear regression analyses. An attempt was made to classify provenances according to their adaptation to the Haney planting site, on the basis of their performance over the years. Within provenance variation for height was compared among provenances and was related to the geography of origin of provenances. Explanations from the point of evolutionary biology were discussed. High values of calculated genetic gain indicated that significant improvement could be achieved by selection. Age - to - age correlations for height were highly significant (p s 0.01) over time and therefore it was concluded that selection is possible after age 7. However, if we consider the importance of the time factor in improvement, programs, it is feasible to select after age 5. Differential results obtained by regression analysis between height and dbh according to zone grouping suggest that volume will be a better criterion for selection purposes than height alone. i i i TABLE OF CONTENTS ABSTRACT i i TABLE OF CONTENTS iv LIST OF TABLES v LIST OF FIGURES v i i ACKNOWLEDGEMENT ix INTRODUCTION 1 History 2 Objectives 3 Experimental Design 3 LITERATURE REVIEW 5 MATERIALS AND METHODS 14 RESULTS AND DISCUSSION 26 Zone Differences 40 Provenance Differences 40 Family Differences 45 Regression Analysis 54 Relationships Among Growth Variables 67 Adaptation to Haney Conditions 79 SUMMARY AND CONCLUSTIONS 84 REFERENCES 86 iv LIST OF TABLES Table 1 Analysis of Variance and Expected Mean Squares for 19 Analysis of Between-Provenance Variation Table 2 Analysis of Variance and Expected Mean Squares for 22 Analysis of Within-Provenance Variation Table 3 Table 4 Least Squares Analysis of Variance for all Seed Zones 27 Means and Standard Deviations of 1986 Total Height, DBH 28 and Volume for all Seed Zones Table 5 Potential Genetic Loss (Reduction from Genetic Gain due to Interactions) 33 Table 6 Additive Genetic Variances (V ) as a Percent of Total 34 A Variance (V /V ) and Narrow Sense Heritabilltles for the A T Traits Studied Table 7 Comparisons of Mean Squares among Zones for all Growth 34 Variables Table 8 Table 9 Components of Variance 37 Total 1986 Height, DBH and Volume Differences Between 41 Seed Zones Table 10 Comparison of Provenance Means in Each Zone with 45 Duncan's Multiple Range Test Table 11 Differences Between the Means of the Best and the 46 Poorest Provenances in all Seed Zones for all Traits Studied Table 12 Differences Between Means of the Best Family in Best 47 Provenance and Poorest Family in Poorest Provenance for all Traits Studied Table 13 Multiple Linear Regression Analysis for Height Considering all Provenances 55 Table 14 Multiple Linear Regression Analysis for DBH Considering all Provenances 56 Table 15 Multiple Linear Regression Analysis for Volume Considering all Provenances 57 Table 16 Multiple Linear Regression for Height by Combinations Methods for Interior Zone all 59 v Table 17 Multiple Linear Regression for Height by all 61 Combinations Method for Zone 1 Table 18 Multiple Linear Regression for Height by all 62 Combinations Method for Zone 3 Table 19 Multiple Linear Regression for Height by all 63 Combinations Method for Zone 2 Table 20 Simple Linear Correlations among Growth and Geographical 65 Variables Table 21 Simple Linear Correlations among Growth Variables 68 Table 22 Juvenile and Mature Correlations of Total Height at 74 Different Years Table 23 Additive Genetic Variance (V ) and (V ) as a Percent of A A Total Phenotypic Variance (v /V ) for Total Height Over Years 76 Table 24 Mean Total Provenance Heights over the Years and Rank- 80 ing of the Provenances vi LIST OF FIGURES Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13 Figure 14 Figure 15 Figure 16 Location of Douglas-fir Provenances IUFRO Plantation, UBC Research Forest, Haney 15 16 Experimental Layout of the Five Seedlings (1 to 5) of 17 the Eight Families Representing a Provenance Block * Provenance Interactions for Height for all 30 Provenances Narrow Sense Heritabilities (Top) and Additive Genetic 35 Variances (Below) Components of Variance as % of Total Variance for all 39 Seed Zones and for all Traits Studied Mean Total Height Differences among Provenances for a l l 42 the Seed Zone Groupings Mean DBH Differences among Provenances for all the Seed 43 Zone Groupings Mean Volume Differences among Provenances for a l l the 44 Seed Zone Groupings Mean Total Height and Volume Differences Between 48 Families of the Best (116) and the Poorest (96) Provenances from Seed Zone 1 Mean Total Height and Volume Differences Between 49 Families of the Best (53) and the Poorest (93) Provenances from Seed Zone 2 Mean Total Height and Volume Differences Between 50 Families of the Best (43) and the Poorest (86) Provenances from Seed Zone 3 Mean Total Height and Volume Differences Between 51 Families of the Best (77) and the Poorest (11) Provenances from the Interior Seed Zone Family Differences of the Best Performing Provenance 52 (43) for all Traits Studied Expected Yield (Volume (m )/ha.) at age 16 53 The Relationship Between Mean Provenance Heights and the 60 Ecophysiological Latitude for Interior Provenances vii Figure 17 The Relationship Between Latitude and Family Variability 70 of Each Provenance for Zone 3 Figure 18 Trends in Components of Variance over the Years for 75 Height Figure 19 Adaptation of Provenances 83 v i i i Acknowledgement The author sincerely wishes to thank the numerous persons who contributed to this thesis: my supervisor, Dr. 0. Sziklai, Professor at the Faculty of Forestry who supplied the data and gave me the opportunity to analyze it; Dr. Judy A. Loo-Dinkins, who discussed with me many of the problems that I encountered; Barry Wong, the former computing officer, who patiently helped me with computers; and Dr. A Kozak who provided me with helpful references and guidance in the statistical part of the study. In addition acknowledgement is extended to the other members of my thesis and examining committees, Dr. D.P. Lavender, Dr. D. Lester, Dr. V. LeMay and Dr. D Tait for contributing their valuable advice and for reviewing the manuscript. Finally I would like to thank my wife, Rosemarie, for editing and typing the manuscript and for her emotional support and encouragement. ix INTRODUCTION Good forest management looks at the genetic quality O f trees and follows certain fundamental principles that will permit the maintenance or even an increase of the quality and quantity of forest production. For this purpose, an understanding of the variability at the species, population or individual level is necessary in order to implement an appropriate improvement program and healthy forest management. As early foresters mentioned, every tree has a different appearance in a naturally regenerated forest. Patterns of this variability differ from region to region; thus different geographical subgroups of trees exhibit different spectra of variation. Maximization of genetic diversity is probably the strategy of nature to optimize the chances that a species can withstand changes in the environment, and differences between populations are the result of a long evolutionary process. Today, deforestation, pollution and climatic change are major threats to the diversity and the collective entity of forest ecosystems. And, because forests are the habitats for diverse organisms, the threat is extended to al l the flora and fauna associated with forests. Ledig (1988) stated the importance of the issue, "Immediate loss of species is not the only danger. Among the surviving species, many populations will be lost, taking with them much of the genetic diversity upon which long-term survival and evolution depend." Increased reforestation programs brought forward a problem of seed movement between localities. Transfer of forest seed from the collection site to another location without considering the adaptation 1 of the seed to the non-native area may prove to be unproductive. It is therefore important to understand variation at the provenance level In order to establish transfer rules and to delineate the limits of seed and planting zones. In other words, the perpetuation of highly resilient, healthy future forest generations depends on our understanding of their diversity and the association of this diversity with different environmental factors. At the moment, means for assessment of genie and geographic variability are biosystematic, genecologic, biochemical and molecular studies. In this study the results of the analysis of data obtained in 1986 from a provenance/progeny trial of Douglas-fir, which Is a type of genecological study, will be presented and these results will be compared to the previous analyses. History: The International Union of Forestry Research Organizations (IUFRO) Section 22 organized seed collections in 1966 and 1968 from the natural range of Douglas-fir to provide material for provenance tests in co-operating countries. The collection of cones was supervised by the Faculty of Forestry, UBC, and as a contribution towards further studies of variability in Douglas-fir on average, fifteen cones per tree from a total of 1818 trees were provided. Yao (1971) studied the biosystematic relationships among families, provenances, sub-regions and regions using thousand-seed weight, (TSW) cone-scale characteristics and germination percent of this material. Differences between and within provenances were revealed at various confidence levels. Therefore UBC maintained the identity of individual 2 trees (families) within each provenance in a nested design in an experiment to test the same material under field conditions, whereas the majority of co-operators established provenance tests from bulked seed. Therefore the identity of the families was not preserved. Seedlings from eight trees per provenance were grown by the Canadian Forest Service (CFS) in 313 styroblock containers in Victoria, 1970. The 1 + 0 containerized seedlings were outplanted at the UBC Research Forest during April, 1971 (Kvestlch, 1976). Objectives: 1 - To estimate the amount of genetic variation for height growth, dbh and volume among and within various provenances, when grown at Haney, British Columbia (B.C.). 2 - To estimate the additive genetic component of variation and narrow sense heritabilities for open-pollinated progenies. 3 - To investigate the juvenile-mature correlations of traits, that is, the consistency of traits over time, in order to recommend the best selection strategy for the Haney planting site. 4 - To investigate the adaptation of provenances to the Haney planting site. Experimental Design: The environmental component in progeny tests is controlled by the use of experimental designs which seek to reduce the non-genetic effects. The use of blocks facilitates control of local variability, exposing the progeny to fairly standard site conditions within a block. Randomization is used to avoid systematic errors in calculation of experimental error. The testing site should be representative of the 3 area where the seed will eventually be used in a normal plantation. The experiment was originally laid out in a randomized complete block design. 4 LITERATURE REVIEW Douglas-fir {Pseudotsuga menziesii [Nirb.] Franco) is one of the most studied species in North America and Europe. Soon after the species was introduced to Europe, the important differences in growth potential and other characteristics between origins of Douglas-fir were recognized. In Livonia, Russia the coastal form was particularly susceptible to cold injury, but the Rocky Mountain forms were inferior in growth rate (Zon, 1913). Isaac (1943) summarized the reproductive habits and s l l v i c s of Douglas-fir as well as other aspects of variability and outlined a proposal for tree improvement research in this species (Isaac, 1949). The typical coast, or green, variety, menziesii, grows better and is commercially more valuable than the other recognized variety, glauca, Rocky Mountain or blue. Fowells tried to delineate the ranges of the two forms separately, and noted that the Rocky Mountain form mixes with the coastal form in southern British Columbia and northeastern Washington (Fowells, 1965). The studies (provenance t r i a l s or common garden methods) of screening geographical variation in Douglas-fir can be reviewed in two phases. The f i r s t phase is the period between 1827, when Douglas f i r s t introduced the species to Europe, and 1966. IUFRO Section 22 started expeditions in 1966 on natural range-wide cone collection and distributed the seed to 59 institutes in 36 countries for provenance research. This marks the beginning of the second phase. Early European experiments were mostly focussed on growth, survival and disease and insect resistance of different origins. Wood (1955) 5 described the general ecological conditions best suited to Douglas-fir and concluded that the "Fraser River type" seed source is the most suitable to use in Britain. In general, the coastal form is better suited to conditions in western Europe and the British Isles than the interior form. One of the oldest extensive studies of thirteen coastal origins in North America was initiated in 1912. Munger authored the plan, and phases of the study have been reported by various authors (U.S.F.S. 1962, 1964). In this study, gene-environment interactions were encountered. The two most Important outcomes of these studies were that, f i r s t , progeny testing at a site different from that to be outplanted is a questionable practice, and secondly, that seed origin is more critical at high elevation than in low elevation plantations. Later, summarizing fifty year results of the same experiment, Sllen (1965) proposed a hypothesis to explain the differing pattern of results of each planting site, based on survival, inherent rates of growth, planting site exposure, weather extremes and time. He stated that "inherent growth rate of a race had developed towards the maximum that could be sustained in each locality against the impacts of long-term weather extremes. Further, these inherent differences in rates of growth for each environment are expressed until climatic extremes intervene at an unpredictable frequency to reduce growth, or damage, or k i l l . " He also entered the time factor to the hypothesis and concluded that, "at the most severe site in this study, both survival and growth rate began to strongly favor the adopted local race by the end of the second decade. At the most sheltered site, where most planted trees 6 have survived, the growth patterns displayed during the first decade were maintained to age 50, with non-local races superior in growth." (Silen, 1965). Sorensen (1967) reported important and rather sharp transitions in seedling growth habits associated with the eastern slopes of the Coast and Cascade Ranges in the Pacific Northwest of the U.S.A. , investigating the west-east transect of the same area. Finally, the changes in genetic parameters over time' were investigated in the same experiment and three periods in the development of genetic variances in height growth were identified. In the juvenile period, variances in environmental error Increased logarithmically while genetic variance within populations existed at moderate levels, and variance among populations was low but increasing; in the early reproductive period, the response to environmental sources of error variance was restricted, genetic variance within populations disappeared, and strong populational differences emerged; in the later period, environmental error again increased rapidly, but genetic variance within populations did not reappear and population differences were maintained at about the same level as established in the early reproductive period (Namkoong et al., 1972). In 1954, the Forest Research Laboratory at Oregon State University and several co-operators initiated a Douglas-fir provenance study based on seed collected from 16 locations throughout the west side of the Cascades in the Pacific Northwest of the U.S.A. Two year-old seedlings of these provenances were outplanted in plantations established at or near each seed collection site in 1959. Twenty and twenty-five year 7 results have been reported (Ching and Hlnz, 1978 and Ching and White, 1985). Significant correlations between growth variables and geographic variables were found, and correlation analyses conducted by individual planting location indicated that different clinal expressions might be exhibited in different plantation environments. The study also drew attention to the observation that local provenances have greatly increased their height rankings as these tests aged. Age-to-age correlations decreased from 0.91 between ages 5 and 9, to 0.48 between ages 5 and 25. They also found that faster growth was positively correlated with early survival before thinnings. In British Columbia, a provenance study employing 16 different coast and interior origins was reported (Haddock et al., 1967). Cold resistance records for two year-old seedlings in the nursery in Vancouver were presented, and the important influence of topography on climate and consequent variation in Douglas-fir was emphasized. The majority of these early provenance experiments employed coastal populations. In 1962 in East Lansing, Michigan, in order to determine the geographic variation patterns of interior populations, a provenance study was established with 128 provenances from the U.S. and Canada. Both east-west and north-south trends in geographic variation were noticed. The lack of elevational trends were explained by continued intermigration which opposed the tendency for genetic differentiation of low- and high-elevation races. Callaham (1964) drew attention to short term studies, emphasizing the value of thorough, biosystematic studies as a prelude to the more expensive field performance trials at the Food and Agriculture 8 Organization (FAO) / IUFRO meeting on Forest Genetics. Short term nursery studies and field experiments were to be developed parallel to the long-term provenance experiments. Some of those early experiments on the physiology of Douglas-fir are worth mentioning. Herman and Lavender (1965) investigated the dormancy period of Douglas-fir and reported that both light intensity and photoperiod affected meristematic activity in shoots and roots until the chilling requirements were satisfied . In another report, they studied the early growth of seedlings from 14 different seedlots from various altitudes and aspects in two different nursery beds as well as in growth rooms. The study provided information on the existence of "aspect races". Greater variation of height growth within seed sources than between seed sources was found and this was attributed to the heterogeneity of the nursery beds (Herman and Lavender, 1967). Irgens-Moller studied the responses of various origins which come from different elevations to temperature and photoperiod under controlled environments and demonstrated the important effect of climate of origin on photoperiod sensitivity as reflected by the date of cessation of height growth (Irgens-Moller, 1957). He also tried to relate the observed differences in growth behaviour under a number of controlled environmental, as well as field, conditions to the differences in environmental conditions of the native habitat. For example, the late onset of dormancy in the seedlings from Vancouver Island was related to the relatively long period favourable for growth in their native habitat. Similarly, the low summer precipitation and short, frost-free season in 1 the northern Rocky Mountains were 9 responsible for the early onset of dormancy in seedlings from this area (Irgens-Moller, 1968) Allen has developed a method to distinguish between coast and interior origins based on seed morphology and germination energy (Allen, 1960). At the beginning of the 1960's with the increasing size of reforestation programs , the supply of sufficient quantities of local seed became a major problem in North America. The only answer to this dilemma was the movement of seed between localities. Haddock (1965) documented the information available for other western species that might have sufficient generality to be applied to the Douglas-fir seed movement problem. Consequently seed collection zones for Douglas-fir in Canada were delineated based on climatic data and the distribution of associated species as well as vegetation maps of the area as a temporary solution to this problem (Haddock and Sziklai, 1966). In 1966, IUFR0 started seed collection expeditions following a survey of the requirements of various member countries and institutes (Fletcher and Barner, 1978). This was the turning point in provenance research in northwest America and Europe. For Douglas-fir by 1970, 326 kilograms of seed comprised of 182 sources were collected and distributed to 59 institutes in 36 countries. This was the first range-wide collection of seed in Douglas-fir. In 1967 Canada participated in the Organization for Economic Co-operation and Development (O.E.C.D). scheme for control of forest reproductive material in international trade and the Canadian Forest Service (C.F.S.) undertook certification of the 1970 seed crop in British Columbia. 10 From August 21 to September 9, 1978, four IUFRO working parties held a joint meeting in Vancouver, B.C. to bring others up to date with the latest results of provenance tests carried out on Douglas-fir, lodgepole pine, Sitka spruce, and Abies provenances. Two volumes of valuable information were gathered. In the f i r s t volume, the environmental characteristics of western North America, distribution, genetics and s i l v i c a l characteristics of the four species, seed procurement problems, and implementation of results from provenance research and f i e l d experiments were reported by various authors from North America and Europe. The performance of the Douglas-fir provenances in various countries, variations in growth characteristics between provenances, correlations to various degrees between growth characteristics and the geographic variables associated with the origin of the provenances were reported and future breeding act i v i t i e s including seed orchard establishment were discussed. Birot proposed a computer based data bank to pool the available data from the experiments established in various countries a l l over the world (Birot, 1978). Campbell reported that source-related variation of Douglas-fir in western Oregon and Washington was mainly clinal (Campbell, 1978). One common feature of these t r i a l s was that the majority of the tests were established from the bulked seed of provenances, therefore intraspecific variation associated with families within provenances remained unknown. Only the French Research Organization and UBC maintained the identity of individual families within provenances. When growing seedlings from 40 northwestern populations of Douglas-fir in eight nursery-bed treatments which contrasted air and 11 s o i l temperatures and nutrition, depending on the trait and the environment in which the test was conducted, i t was found that 20 to 75 % of the source related variance was c l i n a l l y associated with latitude, elevation and distance from the _ocean of parent trees. Population samples interacted significantly with so i l temperature for growth t r a i t s , and with so i l and air temperatures combined for phenological t r a i t s . It was concluded that there is more risk within the northwestern U.S. in moving provenances east-west than north--south, that this risk increases with elevation of provenances and that north-south transfers are more c r i t i c a l near the coast than inland (Campbell and Sorensen, 1978). The genetic variation associated with 193 parent trees for 16 t r a i t s , sampling 115 locations in a 6100 hectare watershed was reported (Campbell, 1979). For most traits the pattern of variation was found to be c l i n a l . The clinal gradient depended upon position within the watershed, and was steeper on the north-facing slope. Also, estimated genotypic values of parent trees differed at identical elevations depending on position in the watershed. The development of electrophoretic techniques during the last two decades provided an alternative to common garden techniques for estimating levels of genetic variation in natural populations (Lewontin, 1974) There are numerous isozymes studies in Douglas-fir (Rudln, 1976; Yeh,1981, Yeh and O'Malloy, 1980; el-Kasaby, 1980). In isozymes studies the amount of genetic variation; whether considering an entire species or a small group of trees, is measured by the average heterozygosity of numerous l o c i . One feature of conifers, that mature seeds contain a 12 haploid gametophyte and a diploid embryo, increases their value in isozyme studies. Enzyme variations at up to 21 loci in natural populations of coastal and interior Douglas-fir from B . C . were studied and results showed that 97% of the total gene diversity resided within populations (Yeh, 1981; Yeh and O'Malley, 1980). However, the study also revealed a high degree of inter-loci variation in heterozygosities within populations. It was concluded that the isozymes surveyed do not seem to be equivalent in their contribution to the overall mean heterozygosity of the populations; therefore many isozyme loci should be surveyed to reliably estimate genetic variation patterns in Douglas-fir. Association between heterozygosity and radial growth rate variables were studied (Mitton, 1981). It was found that the level of heterozygosity was associated with growth variab i l i t y , but the direction of the relationship found in lodgepole pine was opposite to those in ponderosa pine and aspen. In lodgepole pine, high heterozygosity was associated with low variability in growth rate. Although they failed to find an association between mean growth rate and heterozygosity in either ponderosa pine or lodgepole pine, the mean growth rate of highly heterozygous clones of aspen were higher than those of predominantly homozygous clones. 13 MATERIALS AND METHODS In this study, 464 open-pollinated progeny of 58 provenances were analyzed in one set of growing conditions. The cones were collected from the natural range of Douglas-fir from British Columbia to California (Figure 1). The test was established as a randomized complete block design using three blocks in the University of British Columbia (U.B.C.) Research Forest in Haney in 1971 (Kvestich, 1976). Provenances were represented by eight families including five seedlings per family in a row plot. The spacing was three feet between seedlings in a row with 12 feet between rows. 1 + 0 plugs were used as planting stock. The planting site had a gentle slope, south facing aspect, was burned before planting and had a site index of 35 m for a reference age of 50 years measured at breast height (Kvestich, 1976) (Figures 2 and 3). Up to 1979 there were no thinnings and mortality was very low, averaging about 9.8% (Fashler, 1979). The f i r s t thinnings were carried out by Haddock in 1979, by removal of suppressed, misshaped and damaged trees, considering spontaneous mortality and endeavouring to obtain the most even distribution possible of the remaining trees (personal communication, Sziklai). S t a t i s t i c a l analyses were carried out for provenances arranged in groups according to four seed zones (three coastal and one interior). Provenance means were compared using Duncan's multiple range test for each zone. The data were unbalanced with up to three trees per family. Therefore, the analysis of variance for each of three growth variables, 14 15 FIGURE 2 IUFRO Plantation, U.B.C. Research Forest / I.U.F.R.O. Plantat ion, U . B . C . Research Forest , Haney. TIMBER ilOCK 3 / 1 1 1 I \ 1 TIMBER j 98 28 4 0 32 61 52 91 83 45 92 BLOCK 1 10 51 108 63 124 111 87 29 20 53 15 95 34 41 23 93 50 2 6 9 31 11 21 38 96 84 8 102 123 4 5 6 100 74 37 72 85 12 110 106 1 8 97 44 78 36 76 90 58 27 9 9 54 109 60 33 103 104 25 117 67 56 13 4 6 49 73 77 43 17 14 86 6 4 114 19 42 70 79 75 105 8 0 55 22 88 107 8 9 116 57 6 9 48 115 82 66 1 6 39 10 12 B I O C K 2 89 57 99 104 83 79 53 6 7 86 10 109 114 44 11 73 46 25 46 66 9 111 91 77 108 67 28 99 51 98 2 o 01 ' 97 11 6 109 72 2 9 V 1 72 89 116 64 80 45 61 42 83 18 105 95 73 1 I A 1 I O 34 41 103 58 123 96 c o 58 70 88 75 123 87 117 f\ 3 9 36 18 8 84 106 110 JL n 117 63 124 ~29 76 87 93 27 i o 1 JC 37 86 55 12 5 6 80 6 A "T 9 7 39 79 124 108 104 95 60 40 111 25 8 8 26 32 52 93 64 52 53 96 110 38 42 13 14 IS 43 4 27 77 23 92 90 106 7 1 1 3 98 57 9 0 51 3 2 55 66 10 A l 76 23 92 56 60 43 16 now FIGURE 3 Experimental Layout of the five Seedlings (1 to 5) of the Eight Families Representing a Provenance Famlly 1 Famlly 2 Famlly 3 Famlly 4 Famlly 5 i r Famlly 6 Y - • + Famlly 7 3 i r s • — Famlly 8 3 4 \ \ \ • | - T -*\—i H 1 diameter at breast height (dbh), total height and volume by seed zone was performed using a computer program called GENLIN, according to the following linear model: Y =u + P + B + (P*B) + F/P + (B*F/P) + E ljkm J J lj k(l) jk(1) m(ljk) where Y = the measurement of the tree in the k^family in 1 Jkm J the j^block in the provenance u = mean of a l l trees over a l l families, blocks and provenances P = provenance effect B = block effect J (PB)^ = provenance and block interaction F/Pk(i) = family within provenance effect (B*F/P)^( i ) = block and family within provenance interaction E = sampling error m(ijk) K & A l l expected mean squares were derived using a random effects model, using SAS, a computer package available from the UBC Computing Centre (Table 1). Assumptions for the analysis of variance, namely homogeneous variances and normal distribution of observations, were checked with a computer package called MIDAS. The normality of distribution of observations was met. Variances were within an acceptable range and therefore assumed homogeneous. Height measurements were taken to the nearest centimeter, and dbh measurements were taken to the nearest tenth of a centimeter, in 1986. The highest order interaction (B*F/P) was found to be significant at 0.01 confidence level. Therefore sampling error, E , and m(ljk) 18 TABLE 1 ANALYSIS OF VARIANCE AND EXPECTED MEAN SQUARES FOR ANALYSIS OF BETWEEN PROVENANCE VARIATION Source D.F. Expected Mean Squares Provenance Block Prov*Block Fam(Prov) Block*F(Prov) p-1 b-1 p(f-l) V + tV + btV + ftV + bftV BF/P V + tV + E BF/P F/P (p-1)(b-1) V + tV + K E BF/P V + tV + btV E BF/P F/P p(b-l)(f-l) V + tV K E BF/P PB ftV + DftV PB B f tV PB Residual f 1 bpS(t -1) ljk k=l where p = number of provenances b = number of blocks f = number of families nested in provenance t = number of trees per family Vp = variance between provenances V = variance between blocks B Vpe = variance due to block * provenance interaction V = variance between families within provenances F/B V = variance due to block * family within prov. BF/P 1 R interaction V = residual error E experimental error, (B * f/P) J } , were kept separate. Additive genetic variances (V ) for each seed zone were calculated A from the following formula: V = 4 V A F/P since variance among outcrossed half-sib family means, (V ) is the covariance of half-sibs ' F/P 19 and estimates one fourth of V (Becker, 1984). A Narrow sense heritabilities were estimated from the components of variance for each seed zone as follows: 4 V F/P (Falconer, 1960) V + V + V F/P B * F / P E The response to selection at the Haney site was estimated from the following formula: R = i * <r * h2 (Falconer, 1960) p where: i is the intensity of selection expressed in standard deviations above the mean of a normal distribution (in our case a selection intensity of only 1 in 5 individuals from each provenance was chosen; therefore the value for i was taken as 1.4) crp is the phenotypic standard deviation for the traits studied and is calculated as the square root of the denominator of the formula for h2 2 h is the narrow sense heritability The effects of family by block interactions on her i t a b i l i t y and genetic gain estimates were evaluated by using the concept discussed by Matheson and Raymond (1986). For this purpose h e r i t a b i l i t i e s and gains for each zone grouping were calculated considering the presence and absence of these interactions. Therefore the difference is the potential loss of genetic gain i f any of these provenances in the experiment is to be used for reforestation in Haney or similar coastal Douglas-fir 20 ecosystems in the Lower Mainland. The standard error of heritability estimates derived from the interclass correlations was computed as follows (Falconer, 1960; Wright, 1976): (1 - d ) (1 + tbd) 2 S.E.h • (tb) (f - 1) / 2 2 where d = 1/4 h t = number of trees within families b = number of blocks , and f = number of families To investigate within-provenance variation for height, analyses of variance were performed for each provenance according to the following model: Y = u + B + F + (BF) + E ljk i j lj k(ij) where: Y = the mean measurement of the k^ *1 tree in the j*"*1 family in ljk the i*"*1 block u = mean of a l l families over a l l blocks B = block effect l F = family effect J (BF)^ = block * family interaction E = experimental error k(lj) v Table 2 shows the expected mean squares considering random effects. The main objective of the analysis of variance at the provenance level was to generate a variable which would show the variability between families in provenances and relate this variability to 21 geographical variables. For this purpose mean squares of families for each provenance were chosen and regressed to the geographical variables. TABLE 2 ANALYSIS OF VARIANCE AND EXPECTED MEAN SQUARES FOR ANALYSIS OF WITHIN PROVENANCE VARIATION Source of Variation Degrees of Freedom Expected Mean Squares (D.F.) (E.M.S.) Block b - 1 Family f - 1 V + tV + btV E BF F Block * Family (b-1) (f-1) V + tV E BF Error bf (t-l) V E where: V = variance between families F V = variance due to block * family interaction BF J V = residual E b = number of blocks f = number of families nested in provenance t = number of trees per family The zone effect was not included in the model used for analysis of variance; therefore "t" tests were carried out to test the significance of the seed zone means. For hypothesis testing, the following formulas were used (Walpole, 1982): (xx -x2) - (ut -M 2 ) t = r s- -X - X 1 2 22 where: s-X - X = S 1 2 p 1 A^7 v n n and s = / s2 (n -1) + s2 (n -1) P J 1 1 2 2 n + n - 2 1 2 This applies under the following conditions: 2 2 n and n < 30 , and population variances cr = <r . 1 2 • ^ K 1 2 v = n + n - 2 (degrees of freedom) 1 2 xt and x2 are the means for the zones which were compared and u2 are unknown population means s- - is the standard error of differences of two means X - X 1 2 n^ and are ther number of provenances in each zone s2 and s2 are sample variances s is the pooled standard deviation p Volume calculations were carried out using Fortran programming language for a l l individual trees in the experiment using volume equations from both the B.C. Forest Service for immature Douglas-fir and from Kovats (1977) particularly developed for provenance and progeny tests. However, the volume equation from the B.C. Forest Service which uses dbh and height, was adopted, because of the limitations of height ranges of Kovats' equation. The estimation of juvenile by mature correlation was done by simple correlations between total heights at various ages using MIDAS. For this purpose, yearly total height measurements of 100 individual trees, 23 representing one tree from five randomly chosen families of 20 randomly selected provenances, were obtained from previously analyzed data by Fashler (Kvestich, 1976; Fashler, 1979). To investigate the effect of original provenance location on growth performance of provenances in Haney, mean values of provenance height, dbh, and volume were related to each of the geographical variables by simple linear regression. Multiple regression analysis by a l l combinations method was used to select the equation that would best describe the growth variables (dbh, height and volume) using the four Independent geographical variables (elevation, latitude, longitude, and ecophysiological latitude) by zone and by pooling provenances for a l l zones. Ecophysiological latitude was calculated according to the formula: where: E = ecophysiological latitude L = latitude H = altitude in meters To f u l f i l l the fi n a l objective of this study which is to select the best performing provenances for the Haney planting site, a s t a t i s t i c a l technique was used to compare the adaptability of provenances on the basis of the performance of their total height growth. The major d i f f i c u l t y when assessing the adaptability of populations or individuals is how to formulate the environment mathematically. If a physical measure of the environment could be found and considered as fixed then the genotypic means at each site could be regressed against the mean value of the environmental effect. Unfortunately the c r i t i c a l factors causing the interactions are very complex and rarely known in 24 forest ecosystems. It has been suggested that the most relevant measure of the environment over the duration of an experiment is the performance of the experiment i t s e l f (Matheson and Raymond, 1986; Finlay and Wilkinson, 1963). Finlay and Wilkinson (1963) developed a regression approach based on the plantation performance to compare the adaptability of barley varieties grown at several sites for several seasons. Since our experiment is not repeated in several sites, the above mentioned technique was adapted with some modifications, .namely provenance by site interactions were omitted. Even in a uniform edaphic environment a considerable degree of general adaptability w i l l be important, because of the marked fluctuation of climatic conditions from year to year. There were also changes in the experiment over time in terms of spacing after thinnings. Therefore the technique developed in this study may show the responses of the phenotypes to the year to year climatic fluctuation and the changing conditions after thinnings. For this purpose the mean heights of a l l provenances at each year provided a numerical grading of years and were used for the evaluation of the environment; in the results section of this study, i t will be referred to as the "plantation line." For each provenance a linear regression of mean total height on the mean total height of a l l provenances for each year was computed. The slopes of each regression line provided a f a i r l y good measure of provenance performance over the years. From the interpretation of slopes, i t was possible to classify the provenances as stable inferior, unstable or progressive. A l l graphs in this study were done with the Tell-A-Graf graphics program. 25 RESULTS AND DISCUSSION Analysis of the variance for total height, dbh and volume for a l l seed zones is summarized in Table 3. Within zone (between provenance) variation was significant at 0.01 confidence level for a l l of the traits studied. The large range in provenance means indicates considerable genetic variation. This high variability in height, DBH and volume suggests that substantial gains can be made by selecting the most desirable provenances (Table 4). Block differences were not significant for Zone 1 and the interior groupings but significant for Zone 2 and Zone 3 groupings at 0.05 confidence level. When we look at the adaptability values of those provenances which showed no block differences, the majority of them had values lower than 1.00 which are classified as stable inferior growing provenances. Their height growth was under the plantation average; apparently they did not respond to the Haney planting conditions very well. A possible interpretation may be that provenances from Zone 1 and the interior were stable growing in different environments but they were inferior in their height growth, therefore did not respond to the blocking in the experiment, nor did they to the Haney planting site . In contrast to the above situation, provenances from Zone 2 and Zone 3 responded well to the Haney planting conditions. Their adaptability values were greater than 1.00. Their height growth was above the plantation average and most of them were classified as progressive provenances according to their adaptability values for height. Block differences for dbh and volume were not significant at 0.05 confidence level for those provenances which originated in Zone 2 26 TABLE 3 LEAST-SQUARES ANALYSIS OF VARIANCE F VALUES Source of DF Height DBH Volume Test Term Variation SEED ZONE 1 Provenance 16 4. ** 39 3. *• 11 3, ** . 18 MSa Block 2 1. 18 N S 0. 26NS 0. 006NS Block * Prov. Block*Prov 32 4. «* 90 3. 75 4. *» ,42 Block*Fam(Prov) Family/Prov 119 1. 2 7 N S 1. 29NS 1, ,23NS Block*Fam(Prov) B * F/Prov 220 2. ** 50 1. SEED *• 55 ZONE 2 1, ** .78 Residual Provenance 11 6. *» 69 4. «» 40 3. »» ,72 MSa Block 2 4. • 21 1. 9 1N S 2. ,82NS Block * Prov. Block*Prov 22 5. »» 62 3. mm 94 4. mm ,83 Block*Fam(Prov) Family/Prov 84 1. * 42 1. 16NS 1. ,18NS Block*Fam(Prov) B * F/Prov 167 2. 35 1. SEED mm 51 ZONE 3 1. mm ,68 Residual Provenance 18 2. * 41 2. * 07 2. ,32* MSa Block 2 4. • 37 1. 43NS 1. ,41NS Block * Prov. Block*Prov 36 5. *• 99 5. ** 64 5. ** 20 Block*Fam(Prov) Family/Prov 133 1. # 34 1. » 37 1. ,23NS Block*Fam(Prov) B * F/Prov 263 2. »* 61 1. mm 70 1. ,91 Residual INTERIOR ZONES Provenance 9 8. ** 52 6. • * 96 7. • * 53 MSa Block 2 1. 74NS 0. 09NS 0. 29NS Block * Prov. Block*Prov 18 5. »* 27 5. »* 12 4. 36 Block*Fam(Prov) Family/Prov 70 1. 09NS 1. 04NS 0. 77NS Block*Fam(Prov) B * F/Prov 139 1. 81 1. 2 7N S 1. * 31 Residual • P-s 0.05 a MS +MS , . /MS +MS (Zar, 1974) P B*F/P BP F/P ** P £ 0.01 NS not significant 27 TABLE 4 MEANS AND STANDARD DEVIATIONS OF 1986 TOTAL HEIGHT, DBH AND VOLUME IN ALL SEED ZONES Seed Zone Prov. No. Mean Total Height(cm) Mean DBH (cm) Mean Volume (m3) Std. dev. Height Std. dev. DBH Std. dev. Volume 23 1019.0 11.79 0, .0532 177 .6 3 . 13 0 .0344 32 1157 .6 12.48 0 .0645 137 .7 2 .94 0 .0346 55 1088 .4 13.66 0, .0699 145 .0 3 .03 0 .0318 90 1047 .8 13.43 0, .0672 198, .4 3 .07 o' .0337 95 865 .3 10.01 0, .0341 203, .3 2, .73 • 0 .0213 96 720 .8 8.35 0, .0219 215, .9 2, .89 0 .0194 1 99 812 .3 8.98 0. 0268 194, .4 2 .74 0 .0187 104 824 .6 9.52 0, .0300 188, .6 2, .72 0 .0204 106 828 . 1 10.67 0. 0372 220, .9 3, .42 0, .0259 108 943 . 1 11.16 0. 0446 187. 4 2, .98 0, .0273 109 1011 .5 11.79 0. 0506 158, .3 2 .58 0 .0261 110 861 .7 12.00 0. ,0482 195, .0 4, .04 0, .0338 111 843 .6 10.72 o; ,0400 245. 8 3. 67 0, .0279 116 1213 .6 13.41 0. ,0767 223. ,0 2. 69 0, .0349 117 1090, . 1 12.84 0. ,0646 185. 7 2. 88 0, .0345 123 740, .8 8.75 0. ,0247 196. ,4 2. ,94 0, .0228 124 779, .4 9.89 0. 0295 116. ,0 2. ,58 0. 0186 Overall mean 937, .96 11.19 0. 0467 241. , 18 3. ,43 0. 0328 12 1150, .0 12. 13 0. 0592 128. 9 2. , 17 0. 0230 51 1125. 3 11.79 0. 0556 163. 2 2. ,20 0. 0255 52 1252. .5 14.30 0. 0876 136. 6 2. 94 0. ,0377 53 1256. 5 14.80 0. 0954 147. 1 3. 32 0. ,0466 67 1197. 9 13.50 0, 0767 210. 5 3. 18 0. ,0392 2 79 1045. ,2 13.34 0. 0659 174. 6 2. 87 0. ,0348 83 1130. ,9 13.40 0. 0703 181. 8 2. 39 0. 0320 87 1155. ,0 13.57 0. 0741 174. 2 2. 60 0. 0335 89 1175. ,2 14.73 0. 0907 201. 5 3. 92 0. 0508 91 1092. 2 12.46 0. 0579 121. 9 1. 72 0. 0192 92 808. 16 9.66 0. 0296 157. 3 2. 66 0. 0200 93 712. 38 8.58 0. 0213 187. 4 2. 17 0. 0143 Overall Mean 1096. 2 12.74 0. 0659 227. 9 3. 26 0. 0393 25 1197. 5 13.49 0. 0762 153. 9 2. 81 0. 0346 27 1131. 1 12.71 0. 0638 127. 97 2. 60 0. 0271 29 1067. 0 11.48 0. 0496 143. 7 1. 98 0. 0194 39 1023. 8 11.54 0. 0506 167. 5 2. 89 0. 0301 40 1037. 0 11.15 0. 0459 169. 2 2. 03 0. 0200 42 1188. 8 14.68 0. 0861 135. 9 3. 07 0. 0356 43 1272. 0 15.57 0. 1033 138. 7 3. 04 0. 0396 3 56 1194. 5 12.32 0. 0649 147. 2 2. 75 0. 0300 57 1059. 2 13.64 0. 0680 162. 5 2. 76 0. 0313 continued 28 TABLE 4 (continued) Seed Prov. Mean Mean Mean Std. Std. Std. Zone No. Total DBH Volume dev. dev. dev. Height(cm) (cm) (m3) Height DBH Volume 58 1141. ,7 12. .88 0 .0674 209. ,7 2. 64 0. ,0332 60 1135. ,7 12. .05 0, .0598 206. ,4 2. 65 0. ,0301 61 1072. ,8 11. .74 0, .0529 163. 3 2. 52 0. ,0256 72 1185. ,7 12. .93 0, .0687 168. 3 2. 05 0. ,0263 73 1086. 5 12. .97 0, ,0632 197. 6 2. 37 0. ,0260 3 76 963. 6 • 10. ,73 0, .0400 189. ,7 2. 23 0. ,0200 86 802. 8 9. ,26 0, .0280 193.3 2. 68 0. ,0188 88 1071. 5 13. ,32 0. .0677 177. 3 3. 08 0. ,0344 97 980. 7 11. ,42 0. .0522 282. 4 3. 63 0. ,0392 98 996. 1 12. ,92 0. .0624 222. 5 3. 70 0. ,0624 Overall Mean 1086. 0 12. ,466 0. .0617 206. 7 3. 05 0. ,0341 6 911. 3 9. 84 0. ,0357 240. 3 2. 70 0. ,0233 10 655. 3 7. , 17 0. ,0139 162. 4 1. 81 0. ,0088 11 503. 9 5. 53 0. .0076 150. 8 2. 04 0. ,0103 18 648. 1 7. 02 0. 0135 171. 3 2. 00 0. 0105 Interior 28 710. 1 7. 24 0. ,0156 173. 2 2. 09 0. 0113 46 563. 7 6. 29 0. ,0099 139. 5 2. 06 0. 0079 64 954. 8 11. 50 0. ,0476 178. 5 3. 26 0. 0283 66 630. 7 7. 94 0. 0162 124. 4 2. 34 0. 0120 77 1032. 5 11. 76 0. 0551 200. 4 3. 54 0. 0347 88 950. 8 11. 22 0. 0445 173. 4 2. 66 0. 0250 Overall Mean 759. 8 8. 59 0. 0262 248. 1 3. 32 0. 0257 and Zone 3. This was due to the higher provenance by block interactions for dbh and volume for those provenances. Block by provenance interactions were significant at 0.01 confidence level for a l l traits in a l l zone groupings (Figure 4). Therefore i t can be said that provenances behave differently in different blocks. The highest interactions for height were obtained for Zone 1 and Zone 3 (transitional coastal zones) and the lowest for Zone 2 (coastal zone which is most adjacent to the coast). Zone 2 provenances showed higher provenance by block interactions for dbh and volume. Therefore their h e r i t a b i l i t i e s decreased substantially for these t r a i t s . 29 FIGURE 4 Block * Provenance Interaction for Height for A l l Provenances MEAN PROVENANCE HEIGHTS PROVENANCE NUMBER Legend A BLOCK1 X BLOCK2 • BLOCK3 30 We know that radial growth is more affected by density relative to height; apparently these coastal provenances have a good built in superiority for height growth compared to interior and transitional groupings but not for dbh and volume. Further regression analysis revealed that for those provenances, correlations between height and dbh were lower compared to the others. If we relate these findings to natural selection in the evolutionary process, in the coast natural selection probably operated on those genotypes which do not have good competing a b i l i t y for light since growth period and water were not the limiting factors. Block by family within provenance interactions were also significant (P £ 0.01) for a l l traits in a l l zone groupings, and families too showed a trend similar to provenances in their response to blocking. Genotype environment (GE) Interaction receives substantial attention from tree breeders because i t affects tree improvement strategy (Campbell and Shelbourne, 1976; Burdon, 1977; Campbell, 1973; Matheson and Raymond 1986). Campbell (1973) reported that genotype environment interactions in Douglas-fir usually occur in spacing t r i a l s that include greatly different plot densities or widely divergent genotypes. In a subsequent paper they discussed how these interactions affected .the breeding strategy (Shelbourne and Campbell, 1976). They suggested that, f i r s t l y environments must be grouped into breeding zones within which there are minimal interactions with sites both at the provenance and family level; secondly, well adapted populations and genotypes must be selected for high and stable performance in the case of these GE interactions. 31 Burdon (1977) introduced the concept of making use of type A and type B genetic correlations, the former being when both traits have been measured on the same individuals and the latter where the two traits are measured on different individuals within genetic groups, to evaluate GE interactions. He suggested that It would be much easier to evaluate each genotype in just a few environments which are already characterized so as to permit satisfactory extrapolation of performance to the remaining environments rather than test a l l those genotypes over a comprehensive range of environments. Matheson and Raymond (1986) working with tropical species reported that s t a t i s t i c a l significance of GE interactions is not enough to make any decision in breeding strategy. These interactions should be evaluated on the basis of their effect on genetic gain. They suggested that the loss of potential genetic gain caused by the interactions would be a more appropriate measure. However they found that the loss of potential gain was less than 5% for most characters in most species surveyed. In the Haney plantation, there was a grass invasion in the early years of the experiment (Kvestich, 1976) and also thinnings made the spacing irregular within plots. In addition, the genetic material used was very divergent. Hopefully these observed interactions were caused by these factors or are the expression of changes in variance among genotypes at different spacings. Otherwise, up to 1.9%, 1.5% and 2.5% potential genetic losses (reduction from the gain) would be expected for height, dbh and volume respectively (Table 5). Table 6 and Figure 5 show the additive genetic variances and narrow 32 TABLE 5 POTENTIAL GENETIC LOSS (REDUCTION FROM GENETIC GAIN DUE TO INTERACTIONS) Assuming No Interaction Considering Interaction Potential Genetic Zone Variable ] b2 Gain8 % Gainb h2 Gain3 % Gainb % Los; 1 Height 0 .42 78.7 8. .4 0 .25 60.9 6. .5 1, .9 1 DBH 0 .29 1.03 9. ,2 0 .23 0.92 8. .2 1 .0 1 Volume 0 .26 0.0081 17. ,4 0 . 19 0.0070 14. ,9 2, .5 2 Height 0 .57 92.2 8. 4 0, .36 73.4 6. ,7 1. 7 2 DBH 0 . 16 0.51 4. 0 0. . 13 0.50 3. ,7 0. .3 2 Volume 0 . 19 0.0069 10. 5 0, . 14 0.0059 9. 0 1. ,5 3 Height 0 .51 85.3 7. 9 0. .30 65.6 6. 0 1. ,9 3 DBH 0 .39 1.18 9. 5 0. .29 1.01 8. 0 1. 5 3 Volume 0 .27 0.0087 14. 1 0. . 19 0.0073 11. 9 2. 2 Int Height 0, . 10 18.4 2. 4 0. ,07 15.3 2. 0 0. 4 Int DBH 0, .03 0.30 3. 5 0. ,03 0. 10 1. 1 0 Int Volume 0 0 0 0 0 0 0 ° in units of variable, cm. , cm and 3 m for height, dbh and vo respectively gain as % of population (zone) mean sense h e r i t a b i l i t i e s calculated for a l l traits for each zone grouping. Among the coastal zones, the highest values for additive genetic variances were attained by Zone 1 and Zone 3 for dbh and volume, and Zone 2 and Zone 3 for height. Franklin (1979) reported that inter-tree competition may be a major causal factor in the behaviour of additive genetic variance with stand development. M Se and ^B„F / p f °r height were lower for for Zone 2 provenances (Table 7). Therefore, these provenance had higher heritability for height compared to the other coastal zones. M S £ (sampling error) shows the variation within plot and 33 TABLE 6 ADDITIVE GENETIC VARIANCES (V ) AND V AS A PERCENT OF TOTAL VARIANCE A A (V / V ) AND NARROW SENSE HERIABILITY FOR THE TRAITS A T Zone Trait V = 4V V / V h2 SE (h2) A F/P A T 1 Height 7589.24 12. .76 0. 25 0. 50 1 DBH 1.87 15. ,60 0. 23 0. 50 1 Volume 0.000131 11. ,88 0. 19 0. 50 2 Height 7604.00 13. ,80 0. 36 0. 49 2 DBH 0.82 7. 44 0. 13 0. 47 2 Volume 0.000127 7. 92 0. 14 0. 47 3 Height 7255.88 16. 48 0. 30 0. 48 3 DBH 1.80 19. 04 0. 29 0. 48 3 Volume 0.000142 12. 08 0. 19 0. 47 Interior Height 1651.68 2. 52 0. 07 0. 47 Interior DBH 0.14 1. 20 ' 0. 03 0. 47 Interior Volume 0.0001 0 0 TABLE 7 COMPARISONS OF MEAN SQUARES AMONG ZONES FOR ALL GROWTH VARIABLES Variable Zone MS . B*P MS „ B»F/P MS F/P MS E Height 1 196570 40114.8 51023.1 16023.8 Height 2 151590 26965.3 38377.4 11469.1 Height 3 194890 32502.5 43394.0 12451.4 Height Int 160760 30491.0 33245.2 16848.8 DBH 1 34.53 9 .21 11.85 5. ,95 DBH 2 29.93 7 .59 8.83 5. 02 DBH 3 40.46 7 . 18 9.86 4. 22 DBH Int 30.38 5 .93 6.17 4. 65 Volume 1 0.0037 0 .0008 0.0010 0. ,0003 Volume 2 0.0053 0 .0011 0.0013 0. ,0007 Volume 3 0.0049 0 .0009 0.0012 0. ,0005 Volume Int 0.0017 0 .0004 0.0003 0. 0003 34 FIGURE 5 Narrow Sense Heritabilities (Above)and Additive Genetic VarlancesfBelow) 0.4 20NE1 Z0NE2 Z0NE3 INTERIOR SEED ZONES 20 Z0NE1 Z0NE2 Z0NE3 INTERIOR ADDITIVE VARIANCES 35 harbours the effect of environment and 3/4 of additive genetic variance according to the half-sib genetic assumption. *^B*F/P *S ^vPe ^ experimental error which shows the responses of families to blocking, and therefore the responses of genotypes to the site. Apparently, these Zone 2 provenances had very well established genetic superiority for height growth at the population level, and therefore had the highest her i t a b i l i t y for this trait among the coastal groupings. The responses of the families of Zone 2 provenances to different blocks were unpredictable in terms of their dbh and volume. This may be due to greater competition among the families within plots for radial growth. As a consequence, they had lower her i t a b l l i t i e s and additive genetic variances for dbh and volume. These results suggest that any selection among these provenances solely on the merit of their height growth could be misleading. Therefore, a selection criterion which incorporates radial growth would be a better approach. Family-within-provenance variation was significant for height in Zone 2 and Zone 3 and for dbh in Zone 3. B * F/P interactions camouflage the differences between families. From examination of Table 7, i t can be seen that although the Zone 1 grouping had the highest MSF/p, these provenances did not show significant family v a r i a b i l i t y due to very high MS interactions. J e B*F/P Table 8 and Figure 6 show a l l the components of variance for a l l the hypothesized sources of variation. The largest variation was due to provenance differences for height, except in Zone 3. In Zone 3, the largest variation was due to sampling error (V ). For dbh and volume, 36 TABLE 8 COMPONENTS OF VARIANCE Height DBH Volume 7. % P»B F/P BF/P P»B F/P BF/P P»B F/P BF/P 18621.9 100.3 10461.5 1897.3 12395.9 16023.8 22805.6 2955.2 8033.5 1901.0 7870.9 11469.1 7064.0 2212.3 10348.7 1814.0 10129.5 12451.4 31.30 0.17 17.58 3.19 20.83 26.93 41.44 5.37 14.60 3.45 14.30 20.84 16.05 5.03 23.50 4.12 23.01 28.29 SEED ZONE 1 2.1633 18.07 0 1.7118 0.4667 1.6745 5.9538 0 14.30 3.89 13.90 49.74 SEED ZONE 2 2.8753 26.19 0.1604 1.4130 0.2043 1.3058 5.0215 1.46 12.87 1.86 11.89 45.73 SEED ZONE 3 1.1371 12.04 0.0511 2.0945 0.4495 1.4935 4.2195 0.54 22.18 4.76 15.81 44.67 0.000225 0 0.000194 0.000033 0.000188 0.000466 0.000169 0.000055 0.000262 0.000032 0.000226 0.000651 0.000169 0.000007 0.000248 0.000036 0.000226 0.000494 continued. 20.40 0 17. 52 2.97 16.99 42.13 23.61 3.44 16.33 1.98 14.05 40.59 14.33 0.56 21.06 3.02 19. 18 41.87 37 TABLE 8 (continued) Height DBH Volume % % 7= INTERIOR ZONES V p 31273.5 47. 95 4 .6583 39 .95 0.000282 39. 53 V B 859.8 1. 31 0 0 0 0 V P*B 8735.7 13. 39 1 .6480 14 .13 0.000086 11. 98 V F/P 412.9 0. 63 0 .0357 0 .30 0 0 V 7096.3 BF/P 10. 88 0 .6670 5 .72 0.000049 6. 80 V e 16848.8 25. 83 4 .6505 39 .89 0.000298 41. 73 the largest variation was due to sampling error (V ) for a l l E zones noted before, sampling error shows the wlthln-plot variation (variation between individuals of a family). This within-plot v a r i a b i l i t y is attributable to the environmental effect, 3/4 of additive genetic variation and the dominant genetic variation in the half-sib progeny experiments. This value did not vary substantially for dbh and volume among the seed zones, but i t was low for height for Zone 2 and Zone 3 provenances. As is known, height is a good measure of fitness compared to dbh since i t is less affected by the density. Zone 2 and Zone 3 provenances come from coastal ecosystems similar to Haney; therefore in the evolutionary process, they are better adapted to these ecosystems. Consequently, their height growth is less affected by the environment of the Haney planting site as compared to the other zone groupings. The largest value of variation, in the case of dbh, was attributable to V in a l l coastal zones. In contrast, variance 38 FIGURE 6 Components of Variance as % of Total Variance for a l l Seed Zones and For A l l Traits Studied COMPONENTS OF VARIANCE A S % OF TOTAL VARIANCE ZONE 1 ZONE 2 o ZONE 3 INTERIOR o4 ^ v / , , | ol ^=«= , ^ ! VAR(P) VAR® VAR(BP) VAR(^ P) VAR(E) VAR(P) VAR(B) VAR(BP) VAR(F/P) VAR(E) COMPONENTS OF VARIANCE Legend A 1972 x 1975 • 1J7JL_;. B 1986 39 attributable to provenance differences were higher than V£ for interior provenances. For these provenances, block by family interactions were very low compared to the coastal provenances; in addition, they did not respond to blocking in the experiment. A possible interpretation is that they were forced to grow in conditions quite different from their natural habitats; therefore small changes in the environment of the experiment probably did not affect their a b i l i t y to express family differences. If the design of the experiment had been single tree plots (which would reveal inter-tree competition better), then these interior provenances would probably have been eliminated long ago as a result of competition for light. Under the circumstances of the experiment, the design, where each provenance is represented by 40 individuals, allowed their survival since they did not have to compete with coastal provenances on an individual basis. Zone differences: Table 9 shows the differences between zone means for a l l t r a i t s . The "t" test results showed that the means of Zone 1, 2, and 3 were not significantly different at 0.05 confidence level. The interior zone mean was significantly different from the means of Zone 2 and Zone 3, but not different from Zone 1 at 0.05 confidence level. This shows that interior provenances did not adapt well to the Haney planting site . Provenances from Zone 1 produced the lowest height, dbh and volume of a l l the coastal provenances. Provenance differences (within zone differences): Significant differences (psO.01) were found within each zone between provenances. Table 4 and Figures 7, 8 and 9 show the 40 TABLE 9 TOTAL 1986 HEIGHT, DBH AND VOLUME DIFFERENCES BETWEEN SEED ZONES Seed Zone Variable Mean Std. dev. Interior Height DBH Volume Height DBH Volume Height DBH Volume Height DBH Volume 937.96 cm. 11.12 cm. 0.0467 m3 1096.2 cm. 12.74 cm. 0.0650 m3 1086.0 cm. 12.46 cm. 0.06168 m3 759.8 cm. 8.59 cm. 0.02624 m3 241.18 cm. 3.43 cm.^  0.0328 m3 227.92 cm. 3.26 cm. 0.0393 m3 206.70 cm. 3.05 cm. 0.0341 m3 248.14 cm. 3.32 cm. 0.02474 m3 differences between provenances within each zone for a l l traits studied. Standard deviations are not shown in the figures since they are presented in the table. Provenance means were compared using Duncan's test (Table 10). As seen in the table, within each seed zone there is s t i l l substantial heterogeneity in terms of the differences between provenances for a l l traits studied. This heterogeneity is higher for Zone 1 and Zone 3 but low for Zone 2 and the Interior grouping. Family variability was also found to be higher for Zone 1 and Zone 3 in the variance analysis at the provenance level. This variability w i l l be discussed later, in terms of i t s relationship with the geography of variables, in the section dealing with regression. For height, three provenances from Washington (43, 52 and 53) surpassed the overall plantation mean by at least 25%. Six provenances surpassed the overall plantation mean by at least 20%; of these, one was from B.C. (25), one from California (116), and four from Washington (42, 56, 67 and 72). 41 FIGURE 7 Mean Total Height Differences among Provenances for a l l the Seed Zone Groupings MEAN HEIGHTS MEAN HEIGHTS OF 17 PROVENANCES FROM ZONE 1 2 200 ^ ^ 0VCR*Ll UC*N 7Z\ m l lull W W W W i U P M i i i i i m j MEAN HEIGHTS OF 12 PROVENANCES FROM ZONE 2 200 0 l l l l i l P % % ^  P P :P PPP P P in I•VFRAII MFAM 1. 12 51 52 53 67 79 83 87 89 91 92 93 < MEAN HEIGHTS OF 19 COASTAL PROVENANCES M F £ A N HEIGHTS OF 10 INTERIOR PROVENANCES 2 FROM ZONE 3 1200 WOO 800' 600 400 200 H i i m m m 4 mm F?L? v t l u u- M t* N I 4. Ii JfL l l l l l i i f -•IE 111 1 UOOH 1200 1000-800 800 200 0 1 VA O V E I U U . UEAN lililll P P P P P P n 6 10 H 18 28 46 64 66 77 80 PROVENANCE NUMBER 42 FIGURE 8 Mean DBH Differences among Prqvenaces for a l l the Seed Zone Groupings MEAN DBH MEAN DBH OF 17 COASTAL PROVENANCES FROM ZONE 1 x CD Q 2 < 20 «-w-M-12 10-8 6-4-2-0 rail i l l 1 i l l l l l l l l l l l i W l iiiJl MEAN DBH OF 19 COASTAL PROVENANCES FROM ZONE 3 1 * I I inpiiiii Pill In ^ 4 * ? ? > ? * * * ,° * « N> ^  ,« » MEAN DBH OF 12 COASTAL PROVENANCES FROM ZONE 2 • f i l l 1 •JIB 12 51 52 53 67 79 83 87 89 91 92 93 MEAN DBH OF 10 INTERIOR PROVENANCES 10 11 18 28 46 64 66 77 80 PROVENANCE NUMBER 43 FIGURE 9 Mean Volume Differences among Provenances for a l l the Seed Zone Groupings MEAN VOLUME ZONE 1 o CD ZD UJ _ J 9 0.11 0.10 0.09-j 0.08 H 0.07 0.06 0.05 H 0.04 H 0.03 0.02-4 o.oi H 0.00 J l l III II E I ZONE 3 o.n o.io o.os 0.08 0.07-o.o«. o.o JH 0.04 0.03 H 0.02 0.01 -i 0.00 I 11 I L i f hi ZONE 2 fa • • I 111111 12 51 52 53 67 79 83 87 89 91 92 93 INTERIOR PROVENANCES Jf S * ^  * *.? 5- 6 * PROVENANCE NUMBER 44 TABLE 10 COMPARISON OF PROVENANCE MEANS IN EACH ZONE WITH DUNCAN'S MULTIPLE RANGE TEST Seed Zone Variable Number of Number of Provenances Homogeneous Studied Subsets P s 0.05 Height 17 9 1 DBH 17 10 Volume 17 8 Height 12 7 2 DBH 12 6 Volume 12 4 Height 19 10 3 DBH 19 8 Volume 19 8 Height 10 6 Interior DBH 10 5 Volume 10 5 Table 11 Illustrates the differences between the best and the poorest provenances from each zone. To show the percent differences between the best and the poorest provenances, the following equation was used: .. differences between the best and the poorest „ % difference = r X 100 poorest Family Differences: Significant family within provenance variation gives the opportunity for selection of the best families in the best provenance. Table 12 shows the differences between the mean of the best family in the best provenance and poorest family in the poorest provenance. Figures 10, 11, 12 and 13 show the differences between families of the 45 TABLE 11 DIFFERENCES BETWEEN THE MEANS OF THE BEST AND THE POOREST PROVENANCES OF ALL SEED ZONES FOR ALL TRAITS STUDIED Seed Zone Variable Mean of Best Mean of Provenance Poorest Prov. Difference Prov No. (cm) Prov No. (cm) (cm) % Height 116 1213.6 96 720.8 492.8 68 1 DBH 55 13.66 96 8.35 5.31 64 Volume 116 0.0767 96 0.02191 0.0541 250 Height 53 1256.5 93 712.4 544. 1 76 2 DBH 53 14.80 93 8.58 6.22 73 Volume 53 0.09541 93 0.02131 0.07411 348 Height 43 1272.0 86 802.8 469.2 58 3 DBH 43 15.57 86 9.26 6.31 68 Volume 43 0.10331 86 0.02801 0.07531 269 Height 77 1032.5 11 503.9 528.6 105 Interior DBH 77 11.76 11 5.53 6.23 113 Volume 77 0.05501 11 0.007641 0.04741 620 cubic meters best and the poorest provenances for each seed zone. Percent differences were calculated using the same equation as noted before. Differences between families are the most marked with volume, then dbh and height, respectively. Graphic representation of this situation is shown in Figure 14. Assuming, for instance, the selection criterion is to choose those families which surpass the overall provenance mean by 10%. For height, no family surpassed the overall provenance mean by 10%; for dbh only one of the eight families (No. 8), and for volume, two of the eight families exceeded the overall provenance mean by at least 10%. Similar results 46 TABLE 12 DIFFERENCES BETWEEN MEANS OF BEST FAMILY IN BEST PROVENANCE AND POOREST FAMILY IN POOREST PROVENANCE FOR ALL TRAITS STUDIED Mean of Best Mean of Poorest Family of Best Family in Poorest Seed Variable Provenance Provenance Difference Zone Prov!Fam. .! Mean Prov!Fam. .! Mean (cm) % No. !No. ! (cm) No. :NO. ! (cm) Height 116 3 1300.3 96 4 561.0 739.3 132 1 DBH 55 1 15.58 96 4 5.67 9.91 175 Volume 116 7 0.1063 96 4 0.00841 0.09791 1165 Height 53 2 1355.8 93 3 592.5 763.3 129 2 DBH 53 4 16.67 93 3 6.85 9.82 143 Volume 53 6 0.11451 93 3 0.01161 0.10291 887 Height 43 5 1366.5 86 8 514.3 852.2 166 3 DBH 43 8 18.48 86 8 5.65 12.83 227 Volume 43 8 0.14131 86 8 0.00691 0.13441 1948 Height 77 3 1076.8 11 4 441.0 635.8 144 Int. DBH 77 3 12.88 11 5 4.38 8.50 194 Volume 77 4 0.07041 11 5 0.004221 0.0662* 1568 Cubic Meters were observed at the zone level. Therefore, correlations between height and dbh affected the volume performance of the genotypes. It could be said that good height performance does not necessarily mean good volume performance which depends upon the a b i l i t y of provenance for radial growth. Finally, the differences between zones, provenances within zones and families within provenances suggest that a substantial increase in yield could be achieved with selection. Figure 15 illustrates the difference in yield/ha when selecting the best family in the best provenance from the appropriate seed zone at age 15. 47 FIGURE 10 Mean Total Height and Volume Differences Between Families of the Best (116) and the Poorest (96) Provenano From Seed Zone 1 48 FIGURE 11 Mean Total Height and Volume Differences Between Families of the Best (53) and the Poorest (93) Provenances From Seed Zone 2 49 FIGURE 12 Mean Total Height and Volume Differences Between Families of the Best (43) and the Poorest (86) Provenances From Seed Zone 3 MEAN FAMILY HEIGHTS OF PROVENANCE 43 MEAN VOLUME OF FAMILIES IN PROVENANCE 43 NUMBERS OF FAMILIES 50 FIGURE 13 Mean Total Between Families of From Height and Volume the Best (77) and the Interior Seed Differences Poorest Provenances Zone 51 FIGURE 14 Family Differences of the Best Performing Provenance (43) for A l l Traits Studied MEAN FAMILY HDGHTS OF PROVENANCE 43 MEAN DBH OF FAMILIES IN PROVENANCE 43 2 3 4 S 6 7 NUMBERS OF FAMIUES 2 3 4 5 6 7 NUMBERS OF FAMIUES MEAN VOLUME OF FAMIUES IN PROVENANCE 43 2 3 4 5 6 7 NUMBERS OF FAMIUES Please note that for selection purposes, volume is the most important t r a i t . 52 FIGURE 15 Expected Yield (Volume - m3/ha) at Age 16 A B C D One cube represents 20 m3 of volume. This figure represents, from left to right, the expected yield at age 16, A - given by the managed yield tables for Douglas-fir at this age, and when selecting seed from, B -the proper seed zone (ie- vi c i n i t y of Haney ), C -the best provenance or, D - the best family, a l l at age 16. Regression Analysis: Influence of geography of origin on provenance performance Tables 13, 14 and 15 show the influence of geographical variables on the variability of mean height, mean dbh and mean volume of provenances. While 54% of height variability between provenances can be explained by four geographical variables, 47% of this variability in height can be explained by longitude alone (Table 13). 49% of variability in dbh can be explained by the four geographical variables; 44% of this variability was attributable to ecophysiological latitude (incorporation of elevation with latitude) and longitude, the latter being more influential (Table 14). Table 15 shows 44% of variability in volume differences between provenances can be explained by the four geographical variables. Longitude and ecophysiological latitude, the former being more important, accounted for 38% of this volume vari a b i l i t y among provenances. These results suggest that longitude is the most important factor in the growth variability among provenances. The results of the provenance experiment of Association-Foret-Cellulose (Michaud, 1985) which used the same IUFRO material showed, that the provenances east of longitude 121 had weak growth. However, further investigation of the effect of geographical variables on height growth vari a b i l i t y for each zone revealed that each zone grouping exhibited a different spectrum of va r i a b i l i t y according to their natural habitat of origin. Provenances belonging to Zone 1 and interior groupings exhibited different results than those belonging to Zone 2 and 3 groupings. 87% 54 TABLE 13 MULTIPLE LINEAR REGRESSION ANALYSIS FOR HEIGHT BY ALL COMBINATIONS METHOD AND CONSIDERING ALL PROVENANCES Partial Correlation Coefficients Variables 1 2 3 4 R2 SE N X X X X 12 3 4 .3178 .3345 .5612 -.3426 .5381 144 80 40 X X X 12 3 -.1279 -.0086 .5629 .4767 151 97 40 X X X 12 4 .2747 .2908 -.3461 .3258 172 50 40 X X X 13 4 -.0578 .5421 -.0792 .4800 151 50 40 X X X 2 3 4 . 1239 .5436 -.1851 .4862 150 59 40 X X 1 2 -.4636 -.2885 .2340 181 37 40 X X 1 3 -.1383 .6113 .4767 149 91 40 X X 1 4 -.0566 -.3443 .2635 177 84 40 X X 2 3 .0537 .6744 .4680 151 15 40 X X 2 4 .1140 -.5026 .2707 176 96 40 X X 3 4 .5420 -.1484 .4782 149 69 40 X 1 -.4056 . 1645 186 91 40 X 2 -.1558 .0243 201. 98 40 X 3 .6830 .4665 149. 36 40 X A -.5110 .2611 175. 77 40 Height R = multiple coefficient of determination X = elevation l X = latitude 2 X3 = longitude X = ecophysiological latitude (latitude + altitude In hm. ) 4 SE = / MS Error 55 TABLE 14 MULTIPLE LINEAR REGRESSION ANALYSIS FOR DBH BY ALL COMBINATIONS METHOD AND CONSIDERING ALL PROVENANCES Partial Correlation Coefficients Depend. Independent Var. Variables SE DBH X X X X 12 3 4 .3026 .2872 .4852 -.3272 .4932 1 85 . 40 X X X 12 3 -.1266 -.1754 .4925 .4324 1 93 40 X X X 12 4 .2738 .2615 -.3396 .3370 2 09 40 X X X 13 4 .1124 .4730 -.2381 .4476 1 91 40 X X X 2 3 4 -.0525 .4706 -.1810 .4421 1 92 40 X X 1 2 -.4323 -.3961 .2506 2 19 40 X X 1 3 -.0533 .5841 .4144 1 94 40 X X 1 4 .0922 -.4465 .2884 2 14 40 X X 2 3 -.1333 .6117 .4232 1 92 40 X X 2 4 -.0378 -.4716 .2833 2 14 40 X X 3 4 .4696 -.2176 .4405 1 89 40 X 1 -.3334 .1112 2 36 40 X 2 -.2799 .0783 2 40 40 X 3 .6424 .4127 1 92 40 X -.5313 .2823 2 12 40 R = multiple coefficient of determination X = elevation l X = latitude 2 X3 = longitude X = ecophysiological latitude (latitude + altitude in hm.) 4 SE = / MS Error 56 TABLE 15 MULTIPLE LINEAR REGRESSION ANALYSIS FOR VOLUME BY ALL COMBINATIONS METHOD Partial Correlation Coefficients Var. Independent Variables 1 2 3 4 R2 SE N X X X X 12 3 4 .2842 .2857 .4421 -.3109 .4352 .01890 40 X X X 12 3 -.1387 -.1000 .4527 .3748 .01961 40 X X X 12 4 .2628 .2655 -.3274 .2980 .02078 40 X X X 13 4 .0209 .4309 -.1617 .3850 .01945 40 X X X 2 3 4 .0370 .4303 -.1901 .3856 .01944 40 X X 1 2 -.4243 -.3192 .2137 .02169 40 X X 1 3 -.1054 .5279 .3685 .01944 40 X X 1 4 .0125 -.3707 .2448 .02126 40 X X 2 3 -.0424 .5790 .3625 .01953 40 X X 2 4 .0413 -.4622 .2459 .02124 40 X X 3 4 .4307 -.1913 .3847 .01919 40 X 1 -.3528 . 1245 .02259 40 X 2 -.2027 .0411 .02364 40 X 3 .6012 .3614 .01929 40 X 4 -.4946 .2447 .02098 40 VOLUME R = multiple coefficient of determination X = elevation l X = latitude 2 Xg = longitude X = ecophysiological latitude (latitude + altitude in hm.) 4 SE = / MS Error 57 and 71% of height variability among provenances for interior and Zone 1 groupings, respectively, were explained by the four geographical variables. For the interior zone 74% of height var i a b i l i t y was explained by ecophysiological latitude alone (Table 16, Figure 16). Longitude and ecophysiological latitude, the former being more influential accounted for 67% of variability in height growth in Zone 1 (Table 17). Zone 2 and 3 provenances exhibited different spectra of var i a b i l i t y in their height growth than Zone 1 and interior zone groupings. 65% and 30% of height variability among provenances of Zone 3 and Zone 2 were explained by the four geographical variables, with 58% and 11% of vari a b i l i t y between provenances attributable to the effect of latitude for Zones 3 and 2, respectively (Tables 18 and 19). Longitude had very l i t t l e (5% for Zone 3) or no (0.5% for Zone 2) effect on height growth varia b i l i t y for these provenances. These results suggest that for Zone 1 and interior provenances longitude of origin had a pronounced influence on height growth performance. Longitude together with ecophysiological latitude was able to explain, on average, 71% of variation in total height between these provenances. For Zone 2 and Zone 3 provenances, the latitude of origin was the major source of variability; however for Zone 2 these four geographical variables were able to explain only 30% of var i a b i l i t y in height growth. This amount is very low compared to the three other zones. Apparently, elevation of origin had very l i t t l e effect on height va r i a b i l i t y of these Zone 2 provenances. For example, the total height growth difference between high elevation (provenance 52) and low 58 TABLE 16 MULTIPLE LINEAR REGRESSION ANALYSIS FOR HEIGHT BY ALL COMBINATIONS METHOD INTERIOR ZONE Partial Correlation Coefficients Depend. Independent Var. Variables SE Height X X X X 12 3 4 .6695 .6555 .4118 -.6884 .8734 89 74 10 X X X 12 3 -.3568 -.7321 .3363 .7595 112 94 10 X X X 12 4 .6293 .6245 -.6619 .8476 89 90 10 X X X 1 3 4 .3211 .3345 -.7562 .7781 108 48 10 X X X 2 3 4 -.2702 .3017 -.4097 .7706 110 28 10 X X 1 2 -.7667 -.7806 .7288 111 03 10 X X 1 3 .0274 .5032 .4817 153 49 10 X X 1 4 . 1421 -.8000 .7501 106 57 10 X X 2 3 -.6846 .7623 .7244 111 93 10 X X 2 4 -.1028 -.7852 .7477 107 09 10 X X 3 4 . 1726 -.7232 .7526 106 04 10 X 1 .3059 10 X 2 .3421 10 X 3 .4812 10 X .7449 10 R = multiple coefficient of determination X = elevation l X = latitude 2 X = longitude 3 X = ecophysiological latitude (latitude + altitude in hm.) 4 SE = / MS Error 59 FIGURE 16 The Relationship Between Mean Provenance Heights and the Ecophysiological Latitude for Interior Provenances 60 TABLE 17 MULTIPLE LINEAR REGRESSION ANALYSIS FOR HEIGHT BY ALL COMBINATIONS METHOD ZONE 1 Partial Correlation Coefficients Var. Variables 1 2 3 4 R^  SE N X X X X 12 3 4 .3538 .3601 .7198 -.3331 .7175 106.79 10 X X X 12 3 .6085 .4988 .7570 .6822 103.39 10 X X X 12 4 .4687 .4854 -.4608 .4137 140.43 10 X X X 1 3 4 -.1525 .7594 .4823 .6754 104.50 10 X X X 2 3 4 . 1681 .7551 .6000 .6770 104.23 10 X X 1 2 .3082 .5004 .2556 146.50 10 X X 1 3 .4182 .7576 .5769 110.44 10 X X 1 4 -.4504 .4771 .2330 148.71 10 X X 2 3 . 1250 .6216 .4953 120.63 10 X X 2 4 .4680 .2941 .2487 147.18 10 X X 3 4 .8090 .5931 .6676 97.89 10 X 1 .0071 10 X 2 . 1775 10 X 3 .4873 10 X 4 .0379 10 Height R = multiple coefficient of determination X = elevation l X = latitude X3 = longitude X = ecophysiological latitude (latitude + altitude in hm.) 4 SE = / MS Error 61 TABLE 18 MULTIPLE LINEAR REGRESSION ANALYSIS FOR HEIGHT BY ALL COMBINATIONS METHOD ZONE 3 Partial Correlation Coefficients Var. Variables 1 2 3 4 R2 SE N X X X X -.3523 12 3 4 .4209 -.2747 .3604 .6433 93. 47 10 X X X -.0116 12 3 .7411 -.1292 .5901 91. 47 10 X X X -.2948 12 4 .4153 .3145 .6244 87. 57 10 X X X -.7307 1 3 4 -.2123 .7234 .5666 94. 07 10 X X X 2 3 4 X X . 0462 1 2 .7498 .7402 -.1001 .0819 .5928 .5832 91. 85. 17 40 10 10 XX -.1869 1 3 .1182 .0908 126. 13 10 X X -.7389 1 4 .7126 .5461 89. 12 10 X X 2 3 .7516 -.1366 .5901 84. 69 10 X X 2 4 .7671 . 1233 .5886 84. 84 10 X X 3 4 .2640 .1137 .0701 127. 56 10 X 1 .0779 10 X 2 .5823 10 X 3 .0579 10 X 4 .0004 10 2 R = multiple coefficient X = elevation l of determination Height X = latitude 2 X3 = longitude 4 SE = ecophysiological latitude (latitude + altitude in hm. ) / MS E r r o r 62 TABLE 19 MULTIPLE LINEAR REGRESSION ANALYSIS FOR HEIGHT BY ALL COMBINATIONS METHOD ZONE 2 Partial Correlation Coefficients Depend. Independent Var. Variables SE Height X X X X 12 3 4 .3560 .3985 -.1471 -.3776 .3047 74 07 10 X X X 12 3 -.2881 .3862 -.1881 . 1891 73 02 10 X X X 12 4 .3765 .4091 -.3933 .2894 68 36 10 X X X 13 4 -.3978 -.1779 .3644 . 1735 73 71- 10 X X X 2 3 4 .4349 -.1961 -.3157 .2038 72 35 10 X X 1 2 -.2391 .3435 . 1594 68 83 10 X X 1 3 -.2048 .0102 .0470 73 29 10 X X 1 4 -.3644 .3234 . 1465 69 35 10 X X 2 3 .3332 -.0907 .1160 70 60 10 X X 2 4 .3983 -.2672 . 1720 68 31 10 X X 3 4 .0481 .1139 .0181 74.39 10 X 1 .0469 10 X 2 . 1084 10 X 3 .0052 10 X A .0159 10 R = multiple coefficient of determination X = elevation l X = latitude 2 X3 = longitude X = ecophysiological latitude (latitude + altitude in hm. ) 4 SE = / MS Error 63 elevation (53) provenances, both of Matlock, Washington, was only 4 cm., whereas the elevational difference was 1250 feet. It would, of course, be very Interesting to know the aspect of the origin of these provenances. As noted in the literature review, earlier studies revealed that aspect is an important factor in the behaviour of provenance growth (Herman and Lavender, 1965). The lack of an elevational trend in variability in interior populations was explained by continued intermigration (Wright et al., 1971). In this study, the influence of elevation on mean height growth was not significant at 0.05 confidence level for a l l zones (Table 20). However the influence of elevation on height growth was more pronounced for interior provenances compared to coastal provenances. This influence increased especially when elevation was incorporated into latitude . As a result, the correlation between ecophysiological latitude and mean height was highly significant (r = -0.86) for interior provenances (Figure 16). The fact that four geographical variables were able to explain only 30% of variation in height growth of Zone 2 provenances suggests that internal control of height growth is stronger compared to external control. Also, the high heritability estimate for height of these provenances seems to support this explanation. However the relationship between environmental influence and the radial growth for these provenances was more complicated. The correlation between height and dbh for these provenances was found to be very low compared to the other zones. In contrast to the findings of Yeh and Heaman (1982) in which a large positive genetic correlation between height performance and 64 TABLE 20 SIMPLE LINEAR CORRELATIONS AMONG GROWTH AND GEOGRAPHICAL VARIABLES (R VALUES) Zone Variables Elevation Longitude Latitude Ecophysical N ( Latitude Height -0.0840 0.6981 0.4213 0.1947 10 dbh 0.0148 0.5597 0.2869 0.2189 10 Volume -0.0242 0.5886 0.3286 0.1996 10 MS_ f am TSW 0.2508 0.6791 -0.2787 ** -0.8099 -0.4122 -0.5127 0.0145 0.3032 10 10 Height -0.2165 0.0723 0.3292 0.1259 10 dbh -0.0212 -0.4058 -0.2300 -0.2030 10 Volume -0.1162 -0.2920 -0.0761 -0.1378 10 MS_ f am -0.2445 -0.2319 -0.0618 -0.2061 10 TSW 0.4269 -0.2198 -0.4763 -0.1379 10 Height -0.2791 0.2406 0.7631* 0.0207 10 dbh -0.0975 0.1315 0.5764 0.1702 10 Volume -0.1125 0.1280 0.6081 0.1650 10 MS_ f am 0.2954 -0.3177 -0.8844 -0.0173 10 TSW 0.0574 -0.0015 -0.5326 -0.1088 10 Height -0.5531 0.6937* -0.5849 ** -0.8631 10 dbh -0.4942 0.6386 -0.6840* -0.8864 10 Volume -0.4674 0.6487 -0.6590 -0.8512 10 MS_ fam 0.3023 -0.0064 -0.2925 * 0.0396 10 TSW -0.0719 0.4674 -0.7045 -0.5652 10 Height -0.4055** 0.6830 -0.1558 ** -0.5110 40 dbh -0.3334 0.6424 -0.2799 *• -0.5313 40 Volume -0.3528 0.6012 -0.2027 « » -0.4946 40 MS_ fam TSW 0.1950 0.3517 -0.0245 0.1573 -0.355 ** -0.6487 -0.0512 -0.1078 40 40 p s 0.05 » « p s 0.01 considering a l l provenances regardless of originating zone 65 diameter was found, and selection based on height to simultaneously improve both traits was proposed, this present study suggests that these correlations between height and dbh are very much dependent upon the natural habitat of origin. As well as primary growth variables such height and dbh, much interest has been shown in using photosynthetic capacity as an index of growth potential of trees. High rates have been confirmed for Eucalyptus, Populus and Pseudotsuga. However, rates vary appreciably among species as well as among varieties,clones and provenances. The rate of photosynthesis of several Pinus banksiana provenances changed with time (Kozlowski, 1979). One provenance with a very high rate in July had one of the lowest rates in November. Provenances with high rates in October and November also had the highest growth rates. In addition, the amount of seed crop is negatively correlated with vegetative growth. This, of course, complicated the relation between photosynthesis rate and wood production. Two different pathways of photosynthesis have also been identified. Most higher plants and almost a l l trees, except some mangroves and a few others, are classified as plants which follow the Calvin Cycle. C3 plants are less well adapted to undergo environmental stress compared to C (crop) plants. Another disadvantage for C plants is that they 4 3 use atmospheric C02 less efficiently than C4 plants, due to the inhibition of RuBP carboxylase by photorespiration (Kozlowski, 1979). It would be very interesting to investigate the relationship between photosynthetic rate and the stomatal conductance at the provenance level and relate the findings to the vegetative growth performance of the 66 provenances on the same experiment. Furthermore, electrophoretic methods when trees reproduce could also be used for further investigation of the genetic variation at the DNA level and the results could be correlated to physiological and phenotypical variables. Relationships among Growth Variables: Investigation of correlations between height and dbh revealed different results when a l l provenances are considered as opposed to individual zone groupings. When a l l provenances are considered in the analysis, the correlation between height and dbh was very significant (r = 0.9615) (Table 21). However, as seen in the table, investigation of the situation of individual zone groupings showed different results. The highest correlation between height and dbh was obtained for the interior zone grouping (r = 0.9767) and the lowest for Zone 2 (r = 0.6164) (Table 21). The transitional zone groupings (Zone 1 and Zone 3) averaged a value in between (r = 0.92). Zone 2 provenances were the best performing provenances in Haney planting conditions in terms of height growth. This shows once more that selection based on height performance could be misleading. Correlations between TSW and the growth variables (height, dbh and volume) were not significant (p £ 0.05), but there was a positive correlation between TSW and the variability between families within provenances (MS^) for Zones 1 and 3, which are the transitional zones between coastal Zone 2 and the interior. As noted in the Materials and Methods section, MS^  shows the variability among families within a provenance and was generated as a result of 58 separate analysis of variance runs for each provenance, so that within provenance variation 67 TABLE 21 SIMPLE LINEAR CORRELATIONS AMONG GROWTH VARIABLES (R VALUES) Zone Variable Height DBH Volume MS, fam TSW N Height 1.0000 0, ** .9329 0. ** .9623 -0, .2016 -0. ,3723 10 DBH 1. 0000 0. # # .9926 -0, .0751 -0. .2448 10 1 Volume MS, f am TSW 1. 0000 -0, 1. .1101 0000 -0. 0. 1. ,2949 ,6803 ,0000 10 10 10 Height 1.0000 0. .6164 0. ,7765* ,9703 -0. .2151 -0. ,5135 10 DBH 1, .0000 0. 0. .0600 0. ,0149 10 2 Volume MS, f am TSW 1. ,0000 0. 1. .0476 ,0000 -0. -0. 1. , 1762 ,0272 ,0000 10 10 10 Height 1.0000 0. ,9146 0. ** 9211 -0. ** ,8694 -0. 5942 10 DBH 1. 0000 0. ** 9911 -0. 7269 -0. 3458 10 3 Volume MS, f am TSW 1. 0000 -0. 1. ,7034 0000 -0. 0. 1. 3422 6896 0000 10 10 10 Height 1.0000 0. ** 9767 0. «* 9775 ** 9862 0. 1378 0. 4748 10 DBH 1. 0000 0. 0. 1062 0. 4965 10 Int Volume MS, fam TSW 1. 0000 0. 1. 0498 0000 0. o. 1. 5482 0963 0000 10 10 10 Height 1.0000 0. 9615 0. ** 9530 w* 9817 -P- 1077 -0. 1600 40 DBH 1. 0000 0. -0. 0649 -0. 0426 40 Al l Volume 1. 0000 -0. 1168 -0. 1012 40 Prov. MS, fam TSW 1. 0000 0. 1. 4393 0000 40 40 *p s 0.05 p £ 0.01 68 can be investigated as independent from the block by family interactions which camouflaged family variability. When family variability among provenances was correlated to geographical variables, the results were interesting from the view point of evolutionary biology. The simple linear correlation between MS^ , and latitude was very high (r = -0.8844), (Figure 17, Table 20) for those provenances which originated from the zone transitional (Zone 3) between the interior and the coast (Zone 2). This relationship was weaker for Zone 1 (r = -0.4213) and the interior zone (r = -0.2925) but almost non-existent for Zone 2 (r = -0.0618). Ledig, in his talk, "Gene Conservation" given at UBC in the f a l l of 1988, mentioned that diversity within a species decreases when one goes towards the north from the equator. The correlation found in this study between latitude and MS^  is in agreement with this statement. However, the lack of this trend for Zone 2 provenances could be explained by natural selection for high diversity due to favourable climate in the evolutionary process. Apparently plants have different strategies for adaptation, and their strategies depend on the environment they are in. According to the theory of "r" and "k" selection, genotypes which allocate a large fraction of their energy resources to reproductive act i v i t i e s ("r" strategists) w i l l be favoured in environments characterized by high density independent mortality. Conversely, populations subjected to high density dependent mortality ("k" types)will be characterized by genotypes which devoted a larger proportion of resources to vegetative rather than reproductive structures. The "k" selected populations, 69 FIGURE 17 The Relationship Between Latitude and Family Variability of Each Provenance for Zone 3 200000 70 w i l l , therefore, be successful competitively but at the expense of rapid population growth (King and Anderson, 1971; Clegg et ai.,1978). The "k" selection therefore, favours those genotypes best able to buffer the effects of high population density. It is known that Douglas-fir colonized northwestern America from south to north after the glaciers retreated. Probably afterwards individuals adapted to lower elevation coastal Douglas-fir ecosystems and populational differences emerged as a result of natural selection, probably "k" selection. Most probably selection operated against those which did not have good competing a b i l i t y for light since the growth period and water were not the limiting factors. If the above mentioned hypothesis is true, isozyme studies should also reveal these genetic variation at the DNA level, but, as noted before, isozyme studies in Douglas-fir revealed that only 3% of variation in Douglas-fir was attributable to the populational differences. However, electrophoretic data obtained from two different species, Drosophila silvestris and D. heteroneura from the geologically new island of Hawaii was remarkably close, yet the two species were morphologically and cytologically very distinct. (Lewontin, 1974). Significant negative correlations were also found between TSW and MSF for the transitional zone groupings (for Zone 1, r = -0.6803 and for Zone 3, r = -0.6896) (Table 21). This relationship did not exist for interior and Zone 2 groupings. When a l l provenances are considered (Table 20), a negative correlation exists between TSW and latitude (r = -0.6487, p s 0.01). The simple linear correlation between TSW and elevation (r = 0.35, p £ 0.05) was significant. This relationship was 71 stronger in Zone 1, which is the second transitional zone (r = 0.6791, p s 0.05). The relationships among elevation, latitude and TSW found in this study is in confirmation with Yao's previous work on the same material (Yao, 1971). These results suggest that are well defined relationships among TSW, latitude, elevation and MS^ ,. Family variability within provenances for height increased with increasing TSW and decreased with Increasing latitude. Age to Age Correlations and Ontogenic Considerations: Many species have useful juvenile characteristics, which disappear with increasing age, which allow them to withstand negative impacts of the environment. For example, shade tolerance in juvenile stages enables young plants to stay alive under a dense forest canopy. The fact that the entire lower trunk portion of an old beech tree carries leaves, while upper leaves are dropped in winter (leaf retention is a juvenile characteristic of a beech which protects the seedling against cold injury) is the demonstration of disappearance of a juvenile character with age in the same tree. In a beech (Fagus silvatica) experimental scions from the upper and lower branches were grafted onto rootstock of Identical origin, growing under similar conditions in two adjacent rows. The juvenile scions, which were taken from basal epicormic branches, flushed earlier. This demonstrates the control that age has on phenological characteristics of trees. In very heavy shade, young beeches become leaf-shedding,i.e., environment causes the disappearance of their juvenile character, leaf 72 retention. In order to study the extent to which shade influences leaf retention, 14 leafless seedlings growing under a 120-year-old beech stand were selected and transplanted with wide spacing to favourable shade conditions (25% of f u l l l i g h t ) . The results anticipated was that they had left the juvenile leafy stage and therefore they would continue leaf-shedding. However, during the period form 1952 to 1958 a l l surviving plants gradually became leaf-retaining. This was interpreted as a demonstration of the a b i l i t y of shade to prolong the juvenile stage (Schaffalitzky de Muckadell, 1962). Since internal and external controls over ontogenic changes in trees have been demonstrated, tree breeders are Interested in knowing the correlations among quantitative and/or qualitative traits in these different stages for early selection purposes. Age to age correlations for height among various years are given in Table 22. As seen, correlations decreased with increasing age. The simple linear correlation coefficient between 1972 and 1986 total height is 0.47. In other words, only 22% of variation in total height performance at age 16 could be explained by the total height at age 2. This figure increases to 37% at age 5 and to 85% at age 9. Age - to -age correlations are given by Fashler (1979) on the same material for each seed zone between ages 1 and 8. Correlations were very similar in a l l zone groupings. Therefore, in this study, investigation of 2 correlations has been done considering a l l provenances. The r value between total height in 1975 and 1978 is 0.75 (Fashler, 1979). These results therefore suggest waiting at least until age 7 before making any selection on the basis of height performance. Fashler (1979) , 73 TABLE 22 JUVENILE BY MATURE CORRELATIONS OF TOTAL HEIGHTS AT DIFFERENT YEARS r VALUES Year 1973 1975 1980 1981 1982 1983 1984 1985 1986 1973 1. ,00 1975 0. ,86 1. .00 1980 0. .60 0, .75 1. ,00 1981 0. ,58 0. .73 0. ,99 1. 00 1982 0. 56 0. .71 o. ,97 0. ,99 1.00 1983 0. 52 0. .66 0. 95 0. 98 0.99 1.00 1984 0. 49 0. ,64 0. 94 0. 96 0.98 0.99 1.00 1985 0. 48 0. ,62 0. 93 0. 95 0.97 0.98 0.99 1.00 1986 0. 47 0. ,61 0. 92 0. 94 0.96 0.98 0.99 0.99 1.00 A l l of the simple correlations are significant at the 0.01 confidence level (DF = 98) analyzing the data up to age 8, found that for selection purposes, waiting until age 5 would be appropriate. It seems that the age for selection increases as the experiment continues. In this study, Investigation of the development of genetic variances in height growth over time revealed that genetic parameter Vp and VF/p changed with the aging of the experiment (Figure 18). The magnitude of these changes varied according to zone groupings. In the early years of the experiment, Vp was greater than V for Zone 1 and the interior zone. There was a reversed relationship between Vp and V for Zone 2, V being greater than V . The difference between V F/P F/P P P and VF/p was not very well defined for the Zone 3 grouping (Fashler, 1979). Over the years VF/p decreased for a l l zone groupings. This decrease was very pronounced for the interior grouping compared to 74 FIGURE 18 Trends In Components of Variance Over the Years for Height COMPONENTS OF VARIANCE AS % OF TOTAL VARIANCE ZONE 1 ZONE 2 COMPONENTS OF VARIANCE Legend & HEIGHT X DBH • VOLUME other zones. As a parallel to this decrease in the family component (V ) of variance, additive genetic variances and narrow sense F/P h e r i t a b i l i t i e s also decreased (Table 23) and the magnitude of this decrease was more pronounced for Zone 2 and the interior zone. The decrease of V£ was also noticed for a l l zones over the years. The magnitude of this decrease again was more pronounced for the interior and Zone 2 groupings. Changes in genetic parameters for height growth over time was also noted by Namkoong et al. (1972) in a 53 year old experiment. As Franklin (1979) said, "We can i l l - a f f o r d to wait for the mature TABLE 23 ADDITIVE GENETIC VARIANCE (V ) AND V AS A PERCENT A A OF TOTAL PHENOTYPIC VARIANCE (V /V ) FOR TOTAL HEIGHT A T OVER THE YEARS Seed Zone Year V =4V V /V (%) h2 (narrow) A F/P A T 1972 41. , 12 23. ,64 0. 36 1 1975 594. .67 25. .44 0. .42 1978 1911. ,97 19. . 12 0. ,33 1986 7589. .24 12. ,78 0. 25 1972 94. 86 43. ,40 0. 52 2 1975 694. 04 28. 52 0. 35 1978 2854. 15 28. ,40 0. 37 1986 7604. 00 14. 60 0. 36 1972 52. 30 26. ,76 0. 42 3 1975 470. 87 20. ,92 0. ,33 1978 2209. 26 25. 12 0. 39 1986 7255. 88 17. 36 0. 30 1972 23. 31 25. 24 0. 39 Interior 1975 392. 19 24. 32 0. 38 1978 1308. 30 17. 52 0. 28 1986 1651. 68 2. 57 0. 07 76 genotype phase to obtain progeny test results to make selections. The need is to hasten the onset of the mature genotypic phase by inducing fast growth at close spacing and perhaps by manipulating other environmental factors." Campbell et al. (1986) investigated the effect of spacing (square spacing ranging from 30 to 90 cm.) on the variance structure in a population of unselected Douglas-fir over a 9 year period and found that close spacing did not shift the genetic variance structure from the juvenile to the mature phase. However, components of variance for female by spacing effect decreased and male by spacing effect increased with the measurement age. Another study in loblolly pine (Pinus taeda L. ), in a 20 year old open-pollinated progeny test, suggested that genetic parameter changes expressed by families were not in direct response to the onset of competition. The same study also revealed significant differences in the competitive a b i l i t i e s and competitive influences among the seedlings of tested families (Tuskan and Buijtenen, 1985). In our study, estimates of the male effect, the dominant effect, the epistatic effect and the interactions between these effects and the blocks are unattainable. These effects are pooled in V . However, 1979 thinnings increased V interactions in Zone 1, Zone 3 and the interior zone (Figure 18). For Zone 2 provenances, provenance * block interactions did not fluctuate appreciably over the years. Block by family / provenance interactions were not significant up to 1978 (Fashler, 1979), but significant increases were noticed in the 1986 analysis. Thinnings and natural mortality removed 40% of the trees 77 from the experiment up to 1979; Fashler (1979) estimated the natural mortality at 8.9°/. Therefore part of the reason for the increase in B*F/P interactions could be explained by the 1978 thinnings which unregulated the spacing in the experiment. However, physiological experiments In forestry and agriculture suggest the existence of developmental changes or meristematic aging in trees and the interplay between this phenomenon and the environment. For example, gibberellic acid can cause an adult ivy plant [Hedera helix) to revert to the juvenile form. The adult ivy has e l l i p t i c leaves and upright shoots as compared to the juvenile characters of climbing shoots and palmately lobed leaves. Marked changes in the concentration of RNA and DNA as the plant matures and Increases in the RNA/DNA ratio when a plant reproduces are also reported. Flowering of olive trees and grape vines can be promoted by spraying the plants with urac i l , xanthine or caffeine. There is a direct correlation between RNA content and protein synthesis (Schaffalitzky de Muckadell, 1979). These studies suggest that there are biochemical changes triggered by internal and external factors during the changes from the juvenile to the adult stage. An understanding of the ontogenic changes in tree species and their biochemical control would be of great value in forestry especially in nursery activities in order to shorten the time needed in an improvement program. Investigation of the development of variance components according to the different zone groupings suggests that native habitat of provenances has an influence. The different environmental triggers of natural habitats of provenances seem to have a differential conditioning 78 effect depending upon the different regions. Therefore, the juvenile and mature relationships might be different according to the natural habitat of the provenance. Adaptability to Haney Conditions A selection based on the 1986 total height performance would not be able to reflect the responses of these provenances to the year to year fluctuating climate. Therefore the selection technique used in this study wil l consider the performance of the provenances over the years. As seen in Table 24, provenance 91 was the f i r s t ranking provenance in 1973 but gradually decreased i t s ranking to 14*"*1 position in 1986. A selection based on 1986 height would choose this provenance as 14t h choice, whereas according to the slope of regression i t would be the 17t nchoice, considering i t s performance for the entire course of the experiment. Conversely, provenance 25, from Squamish, gradually increased i t s total height ranking from .14^ in 1973 to 4^° in 1986, and therefore would be our third choice. Figure 19 illustrates provenance performance over the years. Height performance of almost a l l interior provenances was significantly under the average plantation performance; they are therefore classified as stable inferior growing provenances. Provenance 11 is. a typical example of such a stable inferior provenance (Table 24, Figure 19). Some provenances, although their height performance is normally under the plantation performance, showed art unusual increase in their ranking in a particular year. Such provenances are classified as unstable provenances. Provenances which attained a cefficient value substantially greater than 1 are classified as progressive provenances; 79 TABLE 24 MEAN TOTAL PROVENANCE HEIGHTS OVER THE YEARS AND RANKING OF PROVENANCES Prov 1973 Rank 1975 Rank 1978 Rank 1986 Rank Reg.0 Rankb No. Ht. Ht. Ht. Ht. (b) Zone 1 23 59, .64 24 138. .0 25 288. .6 26 1019. ,0 24 1. .0428 24 32 59, .90 23 153, .0 19 345, .3 12 1157. ,6 7 1. . 1901 6 55 57. .32 26 140. ,4 24 319, .5 20 1088. ,4 16 1, . 1901 15 90 54. ,48 27 131. .8 26 297. .8 25 1047. ,8 20 1, .0819 20 95 44. ,49 32 111. ,0 33 327. ,7 17 865. ,3 30 0. .8822 30 96 45. ,94 30 99. ,5 35 372. , 1 7 720. ,8 35 0. .7136 35 99 44. ,38 33 110. ,0 34 281. ,8 28 812. ,3 33 0. .8274 33 104 46. ,20 29 111. , 1 32 211. .3 36 824. ,6 32 0, .8493 32 111 40. ,22 35 99. ,5 36 227. .2 34 843. 6 31 0. .8778 31 117 61. ,31 21 144. 5 21 211. .8 35 1090. , 1 15 1. . 1325 14 Zone 2 12 65. 78 18 154. 9 16 313. ,0 22 1150. 0 9 1. , 1813 8 51 75. 09 6 165. 7 11 335. ,7 14 1125. 3 13 1. , 1396 12 52 65. 77 19 173. 9 4 380. 4 5 1252. 5 2 1. ,2826 2 53 69. 47 12 170. 1 8 387. 7 1 1256. 5 1 1. 2857 1 67 78. 88 2 180. 9 1 377. 6 6 1197. 9 3 1. 2091 4 79 74. 34 7 155. 3 15 323. 1 19 1045. 2 21 1. 0537 22 83 77. 96 4 173. 0 6 358. 2 11 1130. 9 12 1. 1374 13 87 71. 05 10 166. 3 10 363. 7 10 1155. 0 8 • 1. 1723 10 89 72. 79 9 179. 3 2 383. 4 3 1175. 2 6 1. 1864 7 91 81. 96 1 177. 6 3 370. 8 8 1092. 2 14 1. 0875 17 Zone 3 25 68. 81 14 171. 8 7 384. 3 4 1197. 5 4 1. 2177 3 27 74. 04 8 168. 3 9 369. 3 9 1131. 1 11 1. 1414 11 29 69. 26 13 154. 1 17 324. 3 18 1067. 0 19 1. 0822 19 40 75. 80 5 158. 3 3 307. 9 23 1037. 0 22 1. 0440 23 continued... 80 TABLE 24 (continued) Prov 1973 Rank 1975 Rank 1978 Rank 1986 Rank Reg.a Rankb No. Ht. Ht. Ht. Ht. (b) Zone 3 42 68. .67 15 173, .2 5 384, .3 2 1188. ,8 5 1. .2071 5 60 60, .63 22 142. . 1 22 299. .2 24 1135. ,7 10 1. . 1752 9 61 78, .36 3 162. .2 12 341. .9 13 1072. ,8 18 1. .0777 21 73 70. . 17 11 156. ,7 14 330. .0 16 1086. 5 17 1. , 1023 16 76 68. .08 16 150. ,6 20 315. .5 21 963. 6 26 0. ,9661 28 86 42. ,84 34 111. ,6 30 237. 4 33 802. 8 34 0. ,8225 34 97 65. , 11 20 153. ,4 18 331. .9 15 980. 7 25 0. ,9843 26 Interior 6 66. 62 17 140. 5 23 273. .9 29 911. 3 29 0. ,9164 29 46 37. ,37 38 89. 5 38 192. 9 37 563. 7 40 0. ,5649 40 64 44. 63 31 114. 2 29 263. ,0 31 954. 8 27 0. ,9925 25 77 39. 61 37 111. 3 31 269. , 1 30 1032. 5 23 1. ,0857 18 80 58. 10 25 128. 6 27 287. 2 27 950. 8 28 0. ,9705 27 10 40. 02 36 96. 4 37 192. 0 38 655. 3 38 0. 6663 37 18 36. 96 39 88. 7 39 183. 6 40 648. 1 37 0. 6642 38 28 46. 74 28 114. 3 28 237. 7 32 710. 1 36 0. 7119 36 66 31. 63 41 80. 7 41 185. 3 39 630. 1 39 0. 6499 39 11 34. 21 40 82. 4 40 164. 7 41 503. 9 41 0. 5045 41 Av.c 59. 1 138. 9 301. 2 982. 3 a regression coefficient (slope) (b) b priority for reforestation c mean total plantation height provenance 53 is such a provenance. The simple linear regression coefficient (slope) for the plantation line is 1.00, because the same values (mean total plantation heights in different years, Table 24, last row) were regressed to each other. 81 Provenance 43, from Marblemount, Washington, a low elevation provenance, which was the best performing provenance according to i t s height and dbh ranking, is not included in the analysis due to the lack of height data in previous years. Provenances 53 and 52 from Matlock, Washington would be selected as f i r s t and second choice according to this analysis. However, local provenances such as provenance 25 from Squamish increased their height rankings over the years. As noted in the literature review, mild climate and sheltered sites may mask the performance of local provenances. Fifty-three year results of an experiment (Silen, 1965) showed that at the most severe site both survival and growth rate began to strongly favour the adapted local race by the end of the second decade. Moving southern provenances northward results in faster growth as a result of a longer growth period, but the trees may then bemore susceptible to f a l l frost damage because they undergo dormancy later. Therefore, to be on the safe side, before making any decision on selection, i t would be appropriate to wait for further results on the same material. 82 FIGURE 19 A D A P T A T I O N O F P R O V E N A N C E S PLANTATION LINE oo CO STABLE INFERIOR o o U J o < CC U l > < o < o CL. 1500 PROGRESSIVE PROVENANCE UNSTABLE fUMtAllOW UNt 1500 PR0VTNANCC AVC.HT.(CM.) • f»0C«H}IVt PB0VO1AMM O mObm,—— IOOO H 1000 500 H 1973 1975 1978 1986 YEAR SUMMARY AND CONCLUSIONS Sixteen-year results are reported for 58 provenances and 464 families of Douglas-fir growing in Haney planting conditions. Results of this analysis confirm those of previous analyses that interior and Zone 1 provenances have slower growth than do the provenances of other zones at Haney conditions. Simple and multiple linear regression analyses conducted by individual zone groupings indicated that different c l i n a l expressions may be exhibited by provenances depending upon the region, or zone grouping, they came from in terms of the effect of geography of origin on provenance performance. Correlations among growth variables (height, dbh, volume, TSW, MS^ ) also varied according to zone groupings. Correlations between height and dbh were lower for Zone 2 provenances, which were the best performing provenances in Haney conditions, compared to the others. Therefore selection based on height alone might be misleading. Ontogenetical changes in gene expression are evident by the changes of variance components over time. Family by block (B*F/P) interactions intensified, while error variance (V£) decreased over the years. Family within provenance variability (V ) decreased and differences between provenances (V ) increased. The magnitude of these changes seems to vary according to zone groupings. However, age - to - age correlations for height indicated significant correspondence over time. Results suggest that early selection is possible after age 7. Because this value was 5 for the previous analysis, therefore i t seems that the age for selection increases as the experiment continues. 84 Regression analyses for height suggest that within provenance varia b i l i t y is not random and probably has an adaptive significance. There are well defined relationships among TSW, latitude, elevation and within provenance variability. Family variability within provenances for height increased with increasing TSW and decreased with increasing latitude. At present, at age 16 the best performing provenances in the experiment are non-local provenances from Washington. However local provenances have greatly increased their height rankings over time. Perhaps, a strong adaptation of Douglas-fir populations to the local environment in which they evolved will become more evident as the experiment approaches the rotation age i f the present trend continues. The s t a t i s t i c a l power of many of the hypotheses tested was probably reduced due to the additional contributions to experimental error(s) by varying family-plot size and early brush invasion in the experiment. The thinnings conducted in 1979 truncated the distribution of trees remaining on a given measurement plot and unregulated the spacing in the experiment. This, therefore may have affected the interactions between families and blocks. Because both spacing and tree size were used as thinning c r i t e r i a , different plot volumes may have been removed thus reducing differences among families and increasing the differences among provenances. 85 REFERENCES Allen, G.S. 1960. A method of distinguishing coastal from interior Douglas-fir seed. B.C. Lumberman 44:26-30. Anderson, R.L. and T.A. Bancroft. 1952. Statistical Theory in Research. McGraw-Hill, New York. 399 pp. Becker, W.A. 1984. Manual of Quantitative Genetics. Academic Enterprises, Pullman, Washington. 175 pp. Birot, Y. 1978. A project of data banking system for the Douglas-fir IUFR0 provenances. In: Proceedings of the IUFR0 Joint Meeting of Working Parties, Vancouver, Canada, Vol. 1:397-408. Burdon R.D. 1977. Genetic correlations as a concept for studying genotype-environment interactions in forest tree breeding. Silvae Genetica 26:168-175. Callaham, Robert Z. 1964. Provenance research: investigation of genetic diversity associated with geography. Unasylva 18:40-50. Campbell, R.K. and B.C. Wilson. 1973. Spacing-genotype interaction in Douglas-fir. Silvae Genetica 22:15-20. Campbell, R.K. 1978. Topoclinal genetic variation in Douglas-fir {Pseudotsuga menziesii var. menziesii). (Abstract) In: Proceedings of the IUFRO Joint Meeting of Working Parties, Vancouver, Canada, Vol. 1:409. Campbell, R.K. and F.C. Sorensen. 1978. Effect of test environment on expression of clines and on delinitation of seed zones in Douglas-fir. Theor. Appl. Genet. 51:233-246. Campbell, R.K. 1979. Genecology of Douglas-fir in a watershed in the Oregon Cascades. Ecology 60:1036-1050. Campbell, R.K.,. R.M. Echols and R.W. Stonecypher. 1986. Genetic variances and interactions In 9-year-old Douglas-fir grown at narrow spacing. Silvae Genetica 35:24-31. Ching, K.K. and P.N. Hinz. 1978. Provenance study of Douglas-fir in the Pacific Northwest region. Silvae Genetica 27:229-233. Ching, K.K. and T.L. White. 1985. Provenance study of Douglas-fir in the Pacific Northwest region. IV. Field performance at age 25 years. Silvae Genetica 34:84-90. Clegg, M.T., A.L. Kohler and R.W. Allard. 1978. Genetic demography of plant populations. In: Ecological Genetics: The Interface. P.F. Brussard (ed). Springer-Verlag, New York, pp 173-187. 86 el-Kassaby, Y.A. 1980. Isozyme patterns of a selected Pseudotsuga menziesii (Mirb.) Franco population. Ph.D. Thesis, UBC, Vancouver. 137 pp. Falconer, D.S. 1981. Introduction to Quantitative Genetics. Longman, London. 340 pp. Fashler, A.M.K. 1979. Intraspecific variation in non-selected natural populations of Douglas-fir {Pseudotsuga menziesii (Mirb.) Franco) M.Sc. Thesis. Faculty of Forestry, UBC, Vancouver, B.C. 74 pp. Finlay, K.W. and G.N. Wilkinson. 1963. The analysis of adaptation in a plant-breeding programme. Australian Journal of Agriculture Research 14:742-754. Fletcher, A.M. and H. Barner. 1978. The procurement of seed for provenance research with particular reference to collection in N. W. America. In: Proceedings of the IUFRO Joint Meeting of Working Parties, Vancouver, Canada, Vol. 1:141-159. Fowells, H.A. 1965. S i l v i c s of Forest Trees of the U.S. Agriculture Handbook No. 271. Franklin, E.C. 1979. Model relating levels of genetic variance to stand development of four North American conifers. Silvae Genetica 28:207-212. Haddock, P.G. 1965. Information available for other western species having sufficient generality to be applied to Douglas-fir seed movement problems. In: Proceeding of Western Forest Genetics Association, Olympia, Washington, December 6-7, 1965. pp. 8-22. Haddock, P.G. and 0. Sziklai. 1966. Seed collection zones for Douglas-fir in Canada. Proceedings of the Sixth World Forestry Congress, Madrid, Vol. 2:1467-1473. Haddock, P.G., J. Walters and A. Kozak. 1967. Growth of coastal and Interior provenances of Douglas-fir (P. menziesii [Mirb] Franco) at Vancouver and Haney. In: B.C. Research Papers. No.79, Faculty of Forestry, UBC, Vancouver, Canada. Herman, R.K. and D.P. Lavender. 1965. Observations on the dormancy period of Douglas-fir (Pseudotsuga menziesii) seedlings. (Abstract) In: Proceedings of Western Forest Genetic Association, Olympia, Washington, p.67. Herman, R.K. and D.P. Lavender. 1967. Early growth of Douglas-fir from various altitudes and aspects in southern Oregon. Silvae Genetica 17:143-151 Irgens-Moller, Holge. 1957. Ecotypic response to temperature and photoperiod in Douglas-fir. Forest Science 3:79-83. 87 Irgens-Moller, Holge. 1968. Geographical variation in growth patterns of Douglas-fir. Silvae Genetica 7:106-110. Isaac, Leo A. 1943. Reproductive Habits of Douglas-fir. Charles Lathrop Park Forestry Foundation, Washington, D.C. 107 pp. Isaac, Leo A. 1949. Better Douglas-fir Forests, from Better Seed. University of Washington Press, Seattle. 64 pp. King, C.E. and W.W. Anderson. 1971. Age-specific selection. II. The interaction between r and k during population growth. The American Naturalist 105:137-156. Kovats, M. 1977. Estimating juvenile tree volumes for provenance and progeny testing. Canadian Journal of Forest Research 7: 335-342. Kvestich, A.M. 1976. Studies in height growth of selected Douglas-fir {Pseudotsuga menziesii (Hirb.) Franco) provenances and progenies. B.Sc. Thesis. Faculty of Forestry, UBC, Vancouver, B.C. 55 pp. Kozlowski, T.T. 1979. Tree Growth and Environmental Stresses. University of Washington Press, Seattle. 192 pp. Ledig, F.T. 1988. The conservation of diversity in forest trees. Why and how should genes be conserved? Bioscience 38:471-478. Lewontin, R.C. 1974. The Genetic Basis of Evolutionary Change. Columbia University Press, New York. 346 pp. Matheson, A.C. and CA. Raymond 1986. A review of provenance*environment interaction: i t s practical importance and use with particular reference to the tropics. Commonwealth Forest Review 65:283-302. Michaud, D. 1985. First results of American Douglas-fir provenance t r i a l s in France. In: Proceedings of of the IUFRO Working Party on Breeding Strategy for Douglas-fir as an Introduced Species, June 1985, Vienna, pp 3-23. Mitton, J.B., P. Knowles, K.B. Sturgeon, Y.B. Linhart and M. Davis. 1979. Associations between heterozygosity and growth rate variables in three western forest trees. In: Proceedings of the Symposium on Isozymes of North American Forest Trees and Forest Insects, USDA, Forest Service, Pacific Southwest Forest and Range Experiment Station. General Technical Report PSW-48. pp 27-33. Namkoong.G., R.A. Usanis and R.R. Silen. 1972. Age-related variation in genetic control of height growth in Douglas-fir. Theoretical and Applied Genetics 42:151-159. Rudin, D. 1976. Biochemical genetics and selection applications of isozymes in tree breeding. In: Proceedings of the IUFRO Joint Meeting on 88 Advanced Generation Breeding, Bordeaux, June 14-18, 1976, pp 145-164. Schaffalitzky de Muckadell, M. 1962. Environmental factors in developmental stages of trees. In: Tree Growth. T.T. Kozlowski (ed) The Ronald Press Company, New York, pp 289-297. Shelbourne, C.J.A. and R.K. Campbell. 1976. The impact • of genotype-environment interactions on tree improvement strategy. In: Proceedings of the IUFRO Joint Meeting of Genetic Working Parties on Advanced Generation Breeding, Bordeaux, pp 73-96. Silen, Roy R. 1965. A 50 year racial study of Douglas-fir in western Oregon and Washington. (Abstract) p.6-7, Western Forest Genetic Association Proceedings, Olympia, Washington. 71 pp. Sorenson, Frank. 1967. Two-year results of a west-east transect-provenance test of Douglas-fir in Oregon. Research Note PNW-72, Pacific Northwest Forest Service, Portland, Oregon, p.1-8. Tuskan, G.A. and J.P. van Buijtenen. 1986. Inherent differences in family response to Inter-family competition in loblolly pine. Silvae Genetica 35:112-118. U.S.F.S. 1962. Annual Report for 1961, Pacific Northwest Forest and Range Experiment Station, U.S.D.A. Forest Service, Portland, Oregon. 105 pp. U.S.F.S. 1964. Annual Report for 1963, Pacific Northwest Forest and Range Experiment Station, U.S.D.A. Forest Service, Portland, Oregon, p 4-6. Walpole, R.E. 1982. Introduction to Statistics. MacMillan Publishing, New York. 521 pp. Wood, R.F. 1955. Studies of northwest American Forests in relation to silviculture. In: Great Britain Forestry Commission Bulletin No. 25, London. 44 pp. Wright, J.W. et al 1971. Genetic variation in Rocky Mountain Douglas-fir. Silvae Genetica 20:54-60. Wright, J.W. 1976. Introduction to Forest Genetics. Academic Press, New York. 463 pp. Yao, C. 1971. Geographic variation in seed weight, some cone scale measurements and seed germination of Douglas-fir {Pseudotsuga Menziesii (Mirb.) Franco). M.Sc. Thesis, UBC. 88 pp. Yeh, F.C.H. and D. O'Malley. 1980. Enzyme variations in natural populations of Douglas-fir (Pseudotsuga Menziesii (Mirb.) Franco) from British Columbia. 1. Genetic variation patterns in coastal populations. Silvae Genetica 29:83-92. 89 Yeh, F.C. 1981. Analyses of gene diversity in some species of conifers In: Proceeding of Symposium on Isozymes of North American Forest Trees and Insects. Gen. Tech. Rep. PSW-48, pp 48-52. Forest Service, USDA, Berkeley, California. Zar, J.H. 1974. Biostatistical Analysis. Prentice-Hall, Inc., Englewood C l i f f s , N.J. Zon, Raphael. 1913. Effects of source of seed upon the growth of Douglas-fir. Forestry Quarterly 11: 499-502. 90 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0075400/manifest

Comment

Related Items