Embed Size (px)
Citation preview
Inheritance of density, microfibril angle, andmodulus of elasticity in juvenile wood of Pinusradiata at two locations in Australia
Brian S. Baltunis, Harry X. Wu, and Mike B. Powell
Abstract: A total of 1640 increment cores from 343 radiata pine (Pinus radiata D. Don) families were sampled at twosecond-generation progeny trials, aged 6 and 7 years, for a detailed genetic study of juvenile wood quality traits. Density,microfibril angle (MFA), and modulus of elasticity (MOE) were determined from pith to bark using SilviScan1 technol-ogy. Heritability was greatest for area-weighted density at the two sites (0.63 and 0.77, respectively), and the lowest forgrowth traits (<0.23). Genotype by environment interaction was low for all three wood quality traits. A positive geneticcorrelation between density and MOE (0.43), and a highly negative, and therefore, favourable genetic correlation betweenMFA and MOE (–0.92) were observed, implying that improvement of multiple juvenile wood properties is possible. Thegenetic correlations between whole-core wood quality traits and individual-ring measurements suggest that improvementfor juvenile wood properties across the entire profile of the corewood including the innermost rings can be achieved. How-ever, density, MFA, and MOE had unfavourable genetic correlations with diameter growth suggesting that selection for in-creased density and MOE, and reduced MFA in the absence of selection for growth will result in a genetic loss for growthrate.
Resume : Au total, 1640 carottes provenant de 343 familles de pin de Monterey (Pinus radiata D. Don) ont ete echantil-lonnees dans deux tests de descendance de seconde generation, ages de 6 et 7 ans, pour faire une etude genetique detailleedes caracteres de qualite du bois juvenile. La densite, l’angle des microfibrilles (AMF) et le module d’elasticite (MOE)ont ete determines de la moelle jusqu’a l’ecorce a l’aide de la technologie SilviScan1. L’heritabilite etait la plus eleveepour la densite ponderee par la surface dans les deux tests (respectivement 0,63 et 0,77) et la plus faible pour les carac-teres de croissance (< 0,23). L’interaction entre le genotype et l’environnement etait faible pour les trois caracteres dequalite du bois. Une correlation genetique positive entre la densite et le MOE (0,43) ainsi qu’une correlation genetique for-tement negative et, par consequent favorable, entre l’AMF et le MOE (–0,92) ont ete observees, ce qui implique quel’amelioration de plusieurs proprietes du bois juvenile est possible. Les correlations genetiques entre les caracteres de l’en-semble du bois de cœur et les mesures de chacun des cernes annuels indiquent que l’amelioration des proprietes du boisjuvenile dans l’ensemble du profil du bois de cœur, incluant les cernes situes le plus loin a l’interieur, est possible. Ce-pendant, les correlations genetiques entre la croissance en diametre et la densite, l’AMF ainsi que le MOE etaient defavor-ables, ce qui indique que la selection pour une densite et un MOE plus eleves et pour un AMF plus faible en l’absence deselection pour la croissance entraınera une perte genetique pour le taux de croissance.
[Traduit par la Redaction]
Introduction
Tree improvement programs have historically placed theirprimary emphasis on improving tree volume and stem form(e.g., stem straightness and branch characteristics). The firstgeneration of improvement of radiata pine (Pinus radiata D.Don) in Australia began in the 1950s. Realized gains in vol-ume after the first generation of breeding of radiata pinewere about 30% over unimproved seedlots (Wright and El-dridge 1985; Matheson et al. 1986), and as of 1990, 100%of the annual planting of radiata pine was from improved se-
lections (Sultech Report 1999). As a result, plantations areproducing merchantable trees at a faster rate with trees har-vested at a younger age.
A major concern with shortening the rotation is that thereis a greater proportion of juvenile wood, which is also calledcorewood (Burdon et al. 2004). Juvenile wood in pines haslower density, thinner cell walls, shorter tracheids, andhigher microfibril angle (MFA) than mature wood (Zobel1981; Megraw 1985; Cown 1992). Lower densities and re-duced fibre dimensions, higher MFA, and low stiffness ofjuvenile radiata pine wood are expected to produce a poorerquality product, often causing dimensional instability (Kib-blewhite and Lloyd 1983; Kretschmann and Bendtsen1992). As a result, one of the main obstacles for greatermarket acceptance of fast grown radiata pine wood is the di-mensional instability of its juvenile core (Cown and vanWyk 2004). For these reasons, wood property traits have be-gun to receive more attention from the radiata pine forest in-dustry and tree improvement programs (Jayawickrama andCarson 2000; Powell et al. 2004).
Received 31 October 2006. Accepted 21 March 2007. Publishedon the NRC Research Press Web site at cjfr.nrc.ca on15 November 2007.
B.S. Baltunis1 and H.X. Wu. Ensis – Genetics, P.O. BoxE4008, Kingston 2604, Australian Capital Territory, Australia.M.B. Powell. Southern Tree Breeding Association, MountGambier, South Australia, Australia.
1Corresponding author (e-mail: [email protected]).
2164
Can. J. For. Res. 37: 2164–2174 (2007) doi:10.1139/X07-061 # 2007 NRC Canada
Wood density, or specific gravity, has often been consid-ered the most important trait in describing wood quality andhas received the most attention (Megraw 1985; Zobel andvan Buijtenen 1989; Burdon and Low 1992). Density canbe measured relatively easily and inexpensively comparedwith other wood quality traits. Although density has beenmost studied and is under strong genetic control comparedwith growth traits (Zobel and van Buijtenen 1989; Burdonand Low 1992; Jayawickrama 2001; Wu et al. 2007), otherwood quality traits such as MFA, spiral grain, and woodstiffness, may also be important in improving overall prod-uct quality in pines. Wood stiffness or modulus of elasticity(MOE) may be the most important wood quality trait forstructural lumber, and is associated with density and MFA.Evans et al. (2000) reported that MFA together with densityaccounted for 94% of the variation associated with MOE inseveral species including radiata pine. Similarly, nearly 93%of the variation in MOE in loblolly pine (Pinus taeda L.) wasaccounted for by MFA and density (Megraw et al. 1999).There are a few genetic studies of clearwood MFA andMOE in radiata pine demonstrating that MFA and MOEare under strong genetic control (Kumar et al. 2002; Kumar2004; Lindstrom et al. 2004; Dungey et al. 2006; Wu etal. 2007). However, these previous studies have used rela-tively small sample sizes or mature trees. Recent advancesin technology such as SilviScan1 (Evans 1994) have al-lowed these wood quality traits to be measured more effi-ciently and to be studied for an entire breeding population.
Juvenile wood can account for as much as 85% of themerchantable volume of 15-year-old loblolly pine, and in30-year-old trees, 30% of the merchantable volume can bejuvenile wood (Zobel and van Buijtenen 1989). Similarly,Cown (1992) reported that approximately 35% of volume ina 25-year-old butt log of radiata pine was juvenile wood,and this proportion increased up to 90% in the uppermostlogs. Any reduction of the juvenile core or genetic improve-ment in the quality of juvenile wood could have broad eco-nomic implications for the forest industry.
The overall focus of our project was to explore the poten-tial for reducing the proportion of juvenile wood (Gapare etal. 2006) and to improve juvenile wood properties in radiatapine by identifying individuals either with less juvenile
wood or with high juvenile density, low MFA, and highstiffness (MOE). The specific objectives of this study wereto (i) determine the genetic variation and inheritance of thethree key juvenile wood property traits (density, MFA, andMOE) in a radiata pine breeding population, (ii) determinethe genetic stability of these traits by estimating the geneticcorrelations across two sites from two major radiata pinegrowing regions, (iii) estimate the genetic correlationsthroughout the profile of the core from pith to bark for eachtrait, and (iv) determine the genetic correlations among den-sity, MFA, MOE, and growth (as measured by ring widthand core length).
Materials and methods
Field trials and genetic materialSince the early 1980s, the Southern Tree Breeding Asso-
ciation (STBA) has been breeding and selecting radiata pinein Australia. The STBA established a series of progeny trialsin 1996 and 1997 from second-generation selections. A totalof about 460 families were planted in 30 progeny trials toform the population for third-generation selections (Powellet al. 2004). Two of these second-generation radiata pineprogeny trials located in two major radiata pine growing re-gions were utilized in this study: BR9611, located at Flynn,Victoria, and managed by Hancock Victorian Plantations PtyLtd, and BR9705, located at Kromelite, South Australia, andmanaged by Green Triangle Forest Products Ltd. Site prepincluded ploughing followed by mounding. The Flynn sitewas fertilized (N/P/K) at a rate of 347 kg/ha in 2000, fol-lowed by a second fertilizer application in 2003 at a rate of329 kg/ha. Trial maintenance at Kromelite included herbi-cide application in the first 2 years of growth aimed at elim-inating competing vegetation. There was no fertilizerapplication at the Kromelite site. These trials contained a to-tal of 343 different families from both full-sib crosses andhalf-sib families from polymix crosses derived from 179 dif-ferent selections in the STBA breeding population. Proge-nies from 43 common parents (including 16 full-sibfamilies) were planted across the two sites.
SamplingA total of 980 trees were sampled from BR9611 (Flynn),
Table 1. Description of two radiata pine second-generation progeny trials and sampling details for ju-venile wood properties.
Sampling detailsTrial BR9611 BR9705
Location Flynn, Victoria, Australia Kromelite, South Australia, AustraliaLatitude 38814’S 37850’SLongitude 146845’E 140855’EElevation (m) 166 55Annual rainfall (mm) 760 900Soil type Sandy loam Sandy clay–loamDate planted June 1996 July 1997Spacing 3.6 m � 2.5 m 2.74 m � 2.5 mNumber of parents 131 91Number of families 249 110Trees sampled per family 4 6Total trees sampled 980 660
Baltunis et al. 2165
# 2007 NRC Canada
while 660 trees were sampled at BR9705 (Kromelite, Ta-ble 1). Two trees from 249 families in each of two replicateswere sampled at BR9611, and two trees from 110 familiesin each of three replicates were sampled at BR9705. Twelvemillimetre bark-to-bark increment cores were collected atbreast height (1.3 m) from these 1640 trees and assessed bySilviScan1. Density was obtained at 50 m intervals, whileMFA was averaged over 5 mm intervals, and these estimateswere used to predict dynamic MOE (Evans 2006). Growthrings were designated as ring number from the pith (pith,ring 1 from pith, ring 2 from pith, etc.). Ring widths weremeasured, basal area was then calculated for each growthring, and measurements of density, MFA, and MOE wereweighted by their individual ring basal areas. In addition,whole-core estimates were determined for all of the growth(core length and area) and wood quality traits.
Statistical analysesAll of the juvenile wood properties (density, MFA, MOE)
and growth (core length, area) traits were analysed in AS-REML (Gilmour et al. 2005) using an individual-tree linearmixed-effects model for both sites individually and jointly.Nonadditive genetic variance was negligible in early runs,and therefore, full-sib family effects were not included inanalyses. Results from single-site analyses were used to ob-tain starting values for the joint-site analyses. Both heteroge-neous additive and error variances were included in themodel. In addition, bivariate analyses were conducted to es-timate the genetic correlation between traits. The followinggeneral model was used to estimate variance componentsand genetic parameters
½1� yi ¼ Xibi þ Ziai þ ei
where yi is the vector of observations indexed (i) by trial inthe case of single-trait analyses across sites or by trait in thecase of bivariate analyses, bi is the vector of fixed effects(i.e., mean, trials, and replicates within trials), and Xi is theknown incidence matrix relating the observations in yi to thefixed effects in bi where
Xibi ¼X1 0
0 X2
� �b1b2
� �
ai is the vector of random genetic effects of individualgenotypes ~MVN(0,G:A) where
G ¼ �2a1
�a1a2
�a1a2 �2a2
� �
and A is the additive genetic numerator relationship matrix,Zi is the known incidence matrix relating observations in yito the genetic effects in ai, �2
a1is the additive genetic var-
iance, �a1a2 is the genetic covariance between additive ef-fects across sites, ei is the random vector of residual terms
� MVN 0I1�
2e1
0
0 I2�2e2
� �� �
�2e1
is the residual variance for each trait, Ii is the identitymatrix of dimension equal to the number of observations ineach trial in the case of analysis of a single trait, and 0 isthe null matrix. When multiple measurements are made onthe same individual then a covariance exists among thesemeasurements. Therefore, in the case of the bivariate ana-lyses, correlated residuals were taken into account.
Estimates of heritability were obtained for each trait ateach site using the variance components from the univariatejoint-site analyses. Standard errors were also estimated usingthe Taylor series expansion method (Kendall and Stuart1963; Namkoong 1979) in ASREML (Gilmour et al. 2005).
The individual-tree narrow-sense heritability for each trait
at each trial (h2
i ) is represented by
½2� h2
i ¼� 2ai
� 2pi
¼� 2ai
� 2aiþ � 2
ei
where �2pi
is the phenotypic variance.To measure the extent of genotype by environment inter-
action for each of the traits, between-site type B genetic cor-relations and their standard errors were estimated using theTaylor series expansion method (Kendall and Stuart 1963;Namkoong 1979) in ASREML (Gilmour et al. 2005).
The type B genetic correlation of additive effects acrosssites is
½3� rBADDITIVE¼ � a1a2ffiffiffiffiffiffiffiffiffiffiffiffiffiffi
� 2a1� 2a2
q
A value of rBADDITIVEclose to one indicates little genotype by
environment interaction, while a low rBADDITIVEindicates ex-
Table 2. Mean, minimum (min) to maximum (max) range, and coefficient of variation (CV) of radiata pine whole-corejuvenile wood properties from trials BR9611 (980 samples) and BR9705 (660 samples).
Trial
BR9611 BR9705Variables Mean Min–max CV (%) Mean Min–max CV (%)
Core length (mm) 75.5 40.3–107.6 13.0 72.6 33.5–97.9 13.6Mean density (kg/m3) 439.2 358.0–542.5 6.3 392.3 325.8–485.5 6.3Mean MFA (8) 32.2 21.6–43.4 10.9 33.8 23.1–45.0 9.2Mean MOE (GPa) 6.1 2.2–10.9 21.1 4.5 2.4–8.2 19.9Area (mm2) 18208 18 208–36 343 25.7 16 857 3 516–30 114 26.1Area Wt. density (kg/m3) 459.7 376.5–551.6 6.4 409 327.6–502.0 6.5Area Wt. MFA (8) 28.6 17.7–43.1 12.8 31.1 20.4–43.2 10.2Area Wt. MOE (GPa) 7.4 2.2–13.4 22.2 5.5 2.4–9.5 19.9
Note: MFA, microfibril angle; MOE, modulus of elasticity.
2166 Can. J. For. Res. Vol. 37, 2007
# 2007 NRC Canada
tensive genotype by environment interaction, with parentalrankings differing between sites.
Similarly, genetic correlations between multiple traitsmeasured on the same individual (type A) were estimatedfrom variance component estimates from the bivariate anal-yses from pooled data across the two sites. Standard errorswere estimated using the Taylor series expansion method(Kendall and Stuart 1963; Namkoong 1979) in ASREML(Gilmour et al. 2005).
The genetic correlation between traits is
½4� rAADDITIVE¼
� axayffiffiffiffiffiffiffiffiffiffiffiffiffi� 2ax� 2ay
q
where �axay is the covariance between additive effects of thetwo traits.
Results and discussion
General trends in growth, density, MFA, and MOEThe mean core length at BR9611 was 75.5 mm with a co-
efficient of variation (CV) of 13.0%, whereas at BR9705(planted 1 year later), the mean core length was 72.6 mmwith a CV of 13.6% (Table 2). Although the total corelength was greater at BR9611, the growth rate was greaterat BR9705 for the first few growth rings (Fig. 1). Only 75%of the sampled trees at BR9611 reached breast height(1.3 m) after 2 years of growth, while 85% of the sampledtrees reached breast height at BR9705 after 2 years ofgrowth indicating that the early growth rate at BR9705 wasfaster, which may be attributed to the weed control at Kro-melite. Additionally, the Kromelite site was observed to bemore uniform at a lower elevation with greater rainfall
BR9705
0
5
10
15
20
25
30
Pith Ring 1 Ring 2 Ring 3 Ring 4 Ring 5 Ring 6
Rin
gW
idth
BR9611
Fig. 1. Width (mm) of individual rings of radiata pine from trials BR9611 and BR9705.
0
100
200
300
400
500
600
Pith Ring 1 Ring 2 Ring 3 Ring 4 Ring 5 Ring 6
Mea
nD
ensi
ty
BR9611
BR9705
Fig. 2. Mean densities (kg/m3) for individual rings of radiata pine from trials BR9611 and BR9705.
Baltunis et al. 2167
# 2007 NRC Canada
(Table 1) and a more productive soil. All of these factorsmay have contributed to faster and more uniform growth atBR9705 than at BR9611.
Overall core density was greater at BR9611 than atBR9705 (Table 2), which is not surprising since fastergrowth rates are typically associated with lower densities inradiata pine. For example, whole-core area-weighted densityat BR9611 ranged from 376.5 to 551.6 kg/m3 with a mean of459.7 kg/m3 (CV = 6.4%). However, at BR9705, the meanarea-weighted density of the whole core was 409 kg/m3 (CV =6.5%) and ranged from 327.6 to 502 kg/m3. Similar trendsare apparent in individual rings with trees at BR9611showing higher density, on average, than trees at BR9705for all rings (Fig. 2). The slight decrease in density in thelast ring was probably due to the fact that tree samplingoccurred during the growing season and may reflect onlyearlywood density in the last ring. However, density ap-
pears to slightly increase with cambial age, which is inagreement with published radiata pine reports (Fig. 2). Forexample, Cown et al. (1992), and Li and Wu (2005) plot-ted density trends indicating this increasing trend in densitywith ring number from the pith. Also, average densitytrends for radiata pine in New Zealand and Australia werereported to range from approximately 400 kg/m3 in rings 1–4 to near 600 kg/m3 in the outer rings (Dungey et al. 2006;Wu et al. 2007) using SilviScan measurements.
Microfibril angle, on the other hand, was comparable atthe two sites. Whole-core area-weighted MFA averaged28.68 (CV = 12.8%) and ranged from 17.78 to 43.18 atBR9611 relative to 31.18 (CV = 10.2%) at BR9705 with asimilar range (Table 2). The profiles of MFA across thewhole core were nearly identical at both sites (Fig. 3). MFAdecreased from approximately 408 in the pith to 208 in rings5 and 6 (Fig. 3). Such a declining trend for MFA from
0
5
10
15
20
25
30
35
40
45
Pith Ring 1 Ring 2 Ring 3 Ring 4 Ring 5 Ring 6
Mea
nM
FA
BR9611
BR9705
Fig. 3. Mean MFAs (8) for individual rings of radiata pine from trials BR9611 and BR9705.
0
2
4
6
8
10
12
14
Pith Ring 1 Ring 2 Ring 3 Ring 4 Ring 5 Ring 6
Mea
nM
OE
BR9611
BR9705
Fig. 4. Mean MOEs (GPa) for individual rings of radiata pine from trials BR9611 and BR9705.
2168 Can. J. For. Res. Vol. 37, 2007
# 2007 NRC Canada
pith to bark was consistent with other radiata pine reports(Donaldson and Burdon 1995; Donaldson 1997; Dungey etal. 2006; Wu et al. 2007). MFA ranged from 308 to 508 inradiata pine corewood, but only ranged from 158 to 258 inthe outerwood (Donaldson 1997), and MFA decreasedmore or less linearly until about age 10 years (Donaldsonand Burdon 1995). Recently, Dungey et al. (2006) and Wuet al. (2007) reported that mean MFA near the pith rangedfrom 358 to 408 and dropped to about 208–308 by aboutring 6 and then continued to decline to 118–138 by rings 11–17 in radiata pine. Similar pith to bark trends in individual-ring values for MFA have also been reported in manyother species, such as loblolly pine (Megraw et al. 1998;Myszewski et al. 2004) and Norway spruce (Picea abies(L.) Karst.) (Herman et al. 1999; Lundgren 2004).
The whole-core MOE was also greater at BR9611 than atBR9705 (Table 2). The mean area-weighted MOE at
BR9611 was 7.4 GPa (CV = 22.2%) and ranged from 2.2 to13.4 GPa, while at BR9705, mean area-weighted MOE was5.5 GPa (CV = 19.9%) and ranged from 2.4 to 9.5 GPa.MOE values reported in the present study showed an in-creasing trend from pith to bark at both sites (Fig. 4). Stiff-ness of the progeny of second-generation selections appearsto be greater than previously reported for a New Zealand ra-diata pine trial, perhaps because of a higher soil fertility thatdrove faster growth rates in the New Zealand trial comparedwith the current study. By a cambial age of 4 years, MOEwas in excess of 8 GPa (Fig. 4), whereas the mean MOE inrings 6–8 of open-pollinated progeny from 72 first-generationselections of radiata pine in New Zealand was 6.3 GPa(Kumar et al. 2002) to 6.6 GPa (Kumar 2004). Dungey etal. (2006) also reported a mean MOE for radiata pine thatwas greater for a 5 � 4 factorial population of radiata pinein Australia than for 50 open-pollinated families in a NewZealand trial.
Heritability and genotype by environment interactionsIndividual-tree narrow-sense heritability was estimated for
all traits at each site. Both unweighted mean data and area-weighted variables were analysed. Narrow-sense heritabilityfor area-weighted and unweighted traits were similar. Forexample, h29611 of unweighted mean density was 0.81 (0.09)as compared to an area-weighted density heritability of 0.77(0.09) (Table 3). Most of the traits had higher heritability es-timates for BR9611 than for BR9705 (Table 3).
As expected, growth traits were less heritable than woodquality traits. The heritability for the growth rate at BR9611was 0.23 (0.07) for both core length and area, while at
BR9705, the h2
9705 was 0.06 (0.07) and 0.09 (0.08) for corelength and area, respectively (Table 3). The heritability forwhole-core area-weighted density was 0.77 (0.09) and 0.63(0.11) for sites BR9611 and BR9705, respectively. Theseheritability estimates for density are in agreement with mostof the existing published estimates for radiata pine. The her-itability for density ranged from 0.53 to 0.96 in several NewZealand radiata pine populations at young and mature ages(Bannister and Vine 1981; Burdon and Low 1992; Kumar2004). However, lower heritabilities have been reported forradiata pine. Li and Wu (2005) reported a lower heritabilityof whole-core area-weighted density of approximately 0.30for 30 open-pollinated families, and this value was stablefrom cambial ages 3 to 26 years. Dungey et al. (2006) re-ported a lower heritability for cumulative area-weighteddensity for a 5 � 4 factorial of radiata pine in Australiawith a maximum heritability of just below 0.40 occurring atring 22.
The heritability of whole-core area-weighted MFA rangedfrom 0.43 (0.11) at BR9705 to 0.63 (0.1) at BR9611, while
the h2
for whole-core area-weighted MOE was 0.36 (0.1)and 0.67 (0.09) at BR9705 and BR9611, respectively(Table 3). The heritabilities of MFA and MOE for BR9705seem lower than BR9611 estimates, but within a similarrange to other radiata pine reports. For example, in clonal
trials of radiata pine, the broad-sense heritabilities (H2) for
MFA were 0.70 (Donaldson and Burdon 1995) and 0.81,
while for MOE, H2
ranged from 0.34 to 0.89 for differentmeasurements of MOE (Lindstrom et al. 2004). Individual-
Table 3. Individual-tree narrow-sense heritability estimates (h2)and the across-site genetic correlation between additive effects(rBADDITIVE
) for radiata pine juvenile wood properties from trialsBR9611 and BR9705.
Variable h29611 rBADDITIVEh29705
Core length 0.23 (0.07) 1.10 (0.57) 0.06 (0.07)Mean density 0.81 (0.09) 0.79 (0.12) 0.64 (0.11)Mean MFA 0.62 (0.10) 0.87 (0.13) 0.37 (0.10)Mean MOE 0.69 (0.09) 0.91 (0.12) 0.37 (0.10)Area 0.23 (0.07) 0.94 (0.45) 0.09 (0.08)Area Wt. density 0.77 (0.09) 0.77 (0.13) 0.63 (0.11)Area Wt. MFA 0.63 (0.10) 0.82 (0.13) 0.43 (0.11)Area Wt. MOE 0.67 (0.09) 0.90 (0.12) 0.36 (0.10)Pith density 0.37 (0.08) 0.92 (0.20) 0.26 (0.10)Pith MFA 0.32 (0.08) 1.02 (0.19) 0.22 (0.09)Pith MOE 0.39 (0.09) 0.45 (0.26) 0.52 (0.11)Pith width 0.02 (0.03) — 0.10 (0.07)Ring 1 density 0.53 (0.11) 0.66 (0.21) 0.55 (0.13)Ring 1 MFA 0.69 (0.10) 0.90 (0.19) 0.25 (0.11)Ring 1 MOE 0.60 (0.10) 0.89 (0.23) 0.23 (0.10)Ring 1 width 0.16 (0.08) — 0.01 (0.07)Ring 2 density 0.55 (0.09) 0.74 (0.16) 0.54 (0.11)Ring 2 MFA 0.52 (0.09) 0.81 (0.17) 0.35 (0.10)Ring 2 MOE 0.60 (0.09) 0.95 (0.12) 0.34 (0.10)Ring 2 width 0.45 (0.11) 0.57 (0.35) 0.19 (0.10)Ring 3 density 0.60 (0.09) 0.75 (0.16) 0.53 (0.11)Ring 3 MFA 0.61 (0.10) 0.76 (0.15) 0.50 (0.11)Ring 3 MOE 0.59 (0.09) 0.87 (0.11) 0.50 (0.11)Ring 3 width 0.20 (0.07) 0.75 (0.26) 0.45 (0.12)Ring 4 density 0.71 (0.09) 0.76 (0.16) 0.37 (0.11)Ring 4 MFA 0.57 (0.09) 0.87 (0.11) 0.57 (0.11)Ring 4 MOE 0.58 (0.09) 0.93 (0.11) 0.41 (0.10)Ring 4 width 0.32 (0.09) 0.37 (0.33) 0.43 (0.12)Ring 5 density 0.47 (0.09) 0.85 (0.24) 0.18 (0.09)Ring 5 MFA 0.53 (0.09) 0.84 (0.11) 0.60 (0.11)Ring 5 MOE 0.51 (0.09) 0.85 (0.12) 0.48 (0.11)Ring 5 width 0.09 (0.05) 0.48 (0.88) 0.05 (0.07)Ring 6 density 0.45 (0.09) — —Ring 6 MFA 0.57 (0.09) — —Ring 6 MOE 0.59 (0.09) — —Ring 6 width 0.12 (0.05) — —
Note: Standard errors are given in parentheses. MFA, microfibril an-gle; MOE, modulus of elasticity.
Baltunis et al. 2169
# 2007 NRC Canada
tree narrow-sense heritability was also reported at nearly0.60 for cumulative area-weighted MOE at ring 6 for asmall radiata pine full-sib family trial in Australia (Dungeyet al. 2006). Wu et al. (2007) reported mean heritabilities of0.79 and 0.50 for MFA and MOE, respectively, from tworotation-aged genetic trials. However, lower values of herit-ability for MOE have been reported in radiata pine (Kumar2004) and for MFA in loblolly pine (Myszewski et al. 2004)and Norway spruce (Hannrup et al. 2004).
Patterns of heritability from pith to bark were also inves-tigated for each trait. Heritability estimates of individualring width were highly variable at both sites (Table 3). Low-est values were observed closest to the pith and heritabilitypeaked between rings 2 and 4. Heritability estimates forindividual-ring density were lower than whole-core esti-
mates of h2. Generally, individual-ring density was more
heritable at BR9611 than at BR9705 as was the case withwhole-core estimates of density (Table 3). At BR9611,
h2
9611 of individual-ring densities ranged from a low of0.37 (0.08) in the pith to a high of 0.71 (0.09) by ring 4,although these estimates were more stable across the coreprofile than for ring width (Table 3). The heritability ofindividual-ring density at BR9705 also increased initiallyfrom the pith to moderately high values (>0.50) throughring 3 before steadily declining in the outermost rings(Table 3).
Within each site, heritability estimates for MFA and MOEon an individual-ring basis from the pith to the outermostrings were very similar (Table 3). At BR9611, for example,heritability was lowest in the pith and then increased to rel-
atively stable levels (h2 � 0:60) across the rest of the core
profile for both MFA and MOE (Table 3). For the mostpart, individual-ring heritabilities for MFA and MOE werelower than whole-core estimates of heritabilities at BR9611(Table 3). The pattern of heritability for MFA and MOEfrom pith to bark at BR9705 was somewhat different. Gen-erally, at BR9705, individual-ring heritabilities for MFA andMOE steadily increased from the innermost to outermostrings (Table 3). The heritability estimates for whole-coreMFA and MOE were generally greater than estimates fromthe innermost rings, but less than estimates from the outerrings at BR9705.
The extent of genotype by environment interaction wasinvestigated for all four traits. Predominantly, there was lit-tle genotype by environment interaction for all of the woodquality traits, both on whole-core and individual-ring meas-urements (Table 3). For instance, for whole-core area-weighted MOE, the type B genetic correlation was 0.90(0.12). The closer rBADDITIVE
is to 1, the better the indicationthat additive effects, or estimated breeding values, shouldrank similarly across sites. If these results for a study at twosites hold for other regions and sites, then forest managerscan have a relatively high level of confidence in making se-lections or deployment decisions for area-weighted MOE injuvenile radiata pine.
Estimates of type B genetic correlations for growth ratewere poorly estimated with large standard errors for whole-core and individual-ring measurements or were much lowercompared with the wood quality traits (Table 3). For example,the type B genetic correlation of core length was 1.10 (0.57).T
able
4.G
enet
ic(a
bove
diag
onal
)an
dph
enot
ypic
(bel
owdi
agon
al)
corr
elat
ions
amon
gra
diat
api
neju
veni
lew
ood
prop
ertie
sfr
omw
hole
-cor
em
easu
rem
ents
from
tria
lsB
R96
11an
dB
R97
05(j
oint
-site
anal
yses
). Cor
ele
ngth
Mea
nde
nsity
Mea
nM
FAM
ean
MO
EA
rea
Are
a-w
eigh
ted
dens
ityA
rea-
wei
ghte
dM
FAA
rea-
wei
ghte
dM
OE
Cor
ele
ngth
–0.5
8(0
.12)
0.25
(0.1
7)–0
.55
(0.1
4)1.
00(0
.001
)–0
.60
(0.1
3)0.
26(0
.16)
–0.5
0(0
.15)
Mea
nde
nsity
–0.1
8(0
.03)
–0.0
8(0
.11)
0.43
(0.0
9)–0
.60
(0.1
2)0.
99(0
.003
)–0
.08
(0.1
1)0.
40(0
.09)
Mea
nM
FA0.
27(0
.03)
–0.0
2(0
.03)
–0.9
0(0
.03)
0.27
(0.1
6)–0
.14
(0.1
1)0.
98(0
.01)
–0.9
0(0
.03)
Mea
nM
OE
–0.3
3(0
.02)
0.37
(0.0
3)–0
.84
(0.0
1)–0
.59
(0.1
4)0.
47(0
.09)
–0.9
0(0
.03)
0.99
(0.0
03)
Are
a0.
99(0
.001
)–0
.19
(0.0
3)0.
27(0
.03)
–0.3
3(0
.02)
–0.6
1(0
.12)
0.28
(0.1
6)–0
.54
(0.1
4)A
rea-
wei
ghte
dde
nsity
–0.1
7(0
.03)
0.97
(0.0
02)
–0.0
7(0
.03)
0.42
(0.0
3)–0
.18
(0.0
3)–0
.14
(0.1
1)0.
43(0
.09)
Are
a-w
eigh
ted
MFA
0.27
(0.0
3)–0
.03
(0.0
3)0.
97(0
.002
)–0
.87
(0.0
1)0.
28(0
.03)
–0.0
8(0
.03)
–0.9
2(0
.02)
Are
a-w
eigh
ted
MO
E–0
.32
(0.0
2)0.
34(0
.03)
–0.8
3(0
.01)
0.98
(0.0
01)
–0.3
2(0
.02)
0.40
(0.0
3)–0
.89
(0.0
1)
Not
e:St
anda
rder
rors
are
give
nin
pare
nthe
ses.
MFA
,m
icro
fibr
ilan
gle;
MO
E,
mod
ulus
ofel
astic
ity.
2170 Can. J. For. Res. Vol. 37, 2007
# 2007 NRC Canada
However, Wu and Matheson (2005) recently reported geno-type by environment interaction for diameter growth in radiatapine, which they attributed to differences in elevation (e.g.,primarily from snow damage at high-elevation sites). Thesetwo sites in the current study were planted at similar elevations,and therefore, elevation was probably not the contributingfactor causing genotype by environment interaction forgrowth. Other possibilities such as soil and nutrition, unequalannual rainfall, drainage, vegetative competition control, pop-ulation sample size, unequal age of samples, or low geneticcontrol could give rise to high or imprecisely quantifiedgenotype by environment interaction for the growth rate.
Genetic correlations among growth, density, MFA, andMOE
The genetic correlation between whole-core mean andarea-weighted measurements for density, MFA, MOE, orgrowth were all near 1 indicating that both unweighted coremean and area-weighted measurements are interchangeable(Table 4).
The genetic correlations between growth and wood qualitytraits were all unfavourable (Tables 4 and 5). The geneticcorrelation between whole-core area-weighted density andcore length was –0.60 (0.13) (Table 4), while the genetic cor-relation between density and ring width ranged from –0.13to –0.78 depending on the ring number from the pith(Table 5). Similar negative genetic correlations were ob-served between area-weighted MOE and core length (–0.50(0.15)) and area (–0.54 (0.14)), whereas annual ring ge-netic correlations between ring width and MOE throughoutthe core ranged from –0.28 to –0.86 (Tables 4 and 5). A
number of studies have reported an unfavourable geneticcorrelation between wood density and growth in radiatapine (Dean et al. 1983; Zobel and van Buijtenen 1989;Dean 1990; Cotterill and Dean 1990; Burdon and Low1992; Jayawickrama 2001; Kumar 2004; Li and Wu2005). Wu et al. (2004) recently reported a mean geneticcorrelation of –0.44 between diameter growth and densityin radiata pine. The estimated genetic correlation betweendensity and growth in the current study was stronger(more negative) than the mean estimated from the earliergenerations in radiata pine. An unfavourable positive ge-netic correlation between area-weighted MFA and growthwas observed, particularly in rings 2–4 from the pith (Ta-bles 4 and 5). Positive, but insignificant, genetic correlationsbetween MFA and growth have also been reported in lo-blolly pine (Myszewski et al. 2004). Unfavourable correla-tions between growth and density, MFA, and MOE implythat selection for increased juvenile wood density, reducedMFA, and increased stiffness in the absence of selection forgrowth rate will all result in decreased diameter growth.
Whole-core area-weighted density and area-weightedMFA had a slight negative genetic correlation, probablystatistically nonsignificant from 0 based on the high standarderror estimate. This is similar to observations by Lindstromet al. (2004) in a 3-year-old radiata pine clonal study and inloblolly pine (Myszewski et al. 2004) and Norway spruce(Hannrup et al. 2004). In contrast, Dungey et al. (2006) re-ported significant and moderately negative genetic correla-tions (–0.60) between density and MFA beginning aroundring 7 and continuing throughout the core profile throughring 25 in radiata pine.
Table 5. Genetic correlations among density, MFA, MOE, and width of individual rings of radiata pinefrom combined analyses of trials BR9611 and BR9705.
Density andMFA
Densityand MOE
MFA andMOE
Density andwidth
MFA andwidth
MOE andwidth
Pith –0.07 (0.16) 0.39 (0.13) –0.81 (0.07) — — —Ring 1 0.13 (0.13) 0.26 (0.13) –0.86 (0.05) –0.13 (0.22) 0.03 (0.22) –0.28 (0.21)Ring 2 –0.05 (0.13) 0.39 (0.11) –0.92 (0.03) –0.34 (0.14) 0.62 (0.11) –0.71 (0.10)Ring 3 –0.11 (0.12) 0.35 (0.11) –0.94 (0.02) –0.67 (0.12) 0.60 (0.11) –0.72 (0.10)Ring 4 –0.28 (0.11) 0.50 (0.09) –0.95 (0.02) –0.45 (0.14) 0.78 (0.09) –0.86 (0.08)Ring 5 –0.23 (0.13) 0.45 (0.11) –0.94 (0.02) –0.78 (0.24) 0.22 (0.24) –0.62 (0.22)Ring 6* –0.34 (0.14) 0.51 (0.11) –0.95 (0.02) –0.20 (0.22) 0.33 (0.19) –0.47 (0.18)
Note: Standard errors are given in parentheses. MFA, microfibril angle; MOE, modulus of elasticity.*Estimates based on analyses involving trial BR9611 only.
Table 6. Genetic (above diagonal) and phenotypic (below diagonal) correlations among radiata pine individual ring density mea-surements and the area-weighted density of the whole core from trials BR9611 and BR9705 (joint-site analyses).
Pith Ring 1 Ring 2 Ring 3 Ring 4 Ring 5 Ring 6* Whole core
Pith 0.91 (0.06) 0.81 (0.07) 0.80 (0.08) 0.77 (0.08) 0.73 (0.09) 0.91 (0.08) 0.84 (0.06)Ring 1 0.41 (0.03) 0.94 (0.03) 0.96 (0.03) 0.80 (0.07) 0.79 (0.07) 0.88 (0.07) 0.96 (0.02)Ring 2 0.36 (0.02) 0.69 (0.02) 0.94 (0.03) 0.87 (0.05) 0.86 (0.05) 0.93 (0.04) 0.97 (0.02)Ring 3 0.33 (0.03) 0.55 (0.02) 0.62 (0.02) 0.93 (0.03) 0.89 (0.05) 0.84 (0.07) 0.99 (0.01)Ring 4 0.29 (0.03) 0.51 (0.02) 0.52 (0.02) 0.66 (0.02) 0.82 (0.06) 0.80 (0.07) 0.94 (0.02)Ring 5 0.21 (0.03) 0.42 (0.03) 0.46 (0.02) 0.47 (0.02) 0.58 (0.02) 0.82 (0.04) 0.90 (0.04)Ring 6* 0.27 (0.03) 0.50 (0.03) 0.51 (0.03) 0.52 (0.03) 0.59 (0.03) 0.69 (0.02) 0.91 (0.04)Whole core 0.42 (0.02) 0.75 (0.01) 0.78 (0.01) 0.84 (0.01) 0.83 (0.01) 0.69 (0.01) 0.76 (0.02)
Note: Standard errors are given in parentheses.*Estimates based on analyses involving trial BR9611 only.
Baltunis et al. 2171
# 2007 NRC Canada
The genetic correlation between whole-core area-weightedMFA and area-weighted MOE was highly negative (–0.92(0.02)) (Table 4), and a continuing negative trend was ob-served throughout the profile of the core with individual-ring genetic correlations ranging from –0.81 to –0.95(Table 5). Similarly, highly negative correlations betweenMFA and MOE were reported by Lindstrom et al. (2004)and Dungey et al. (2006) in radiata pine. A strong geneticcorrelation may imply that the same genes may be responsi-ble for the two traits (pleiotropy). This strong negative rela-tionship indicates that selection for reduced MFA would leadto gains in MOE throughout the juvenile core or vice versa.
The genetic correlation between MOE and density wasalso favourable indicating that selection for density shouldalso lead to improvement in overall stiffness for juvenilewood or vice versa (Tables 4 and 5). The genetic correlationwas moderately positive between whole-core area-weighteddensity and MOE (0.43 (0.09)). Similarly, the genetic corre-lation between density and MOE across the profile of thecore from pith to bark was also positive and ranged from0.26 to 0.51 (Table 5). Kumar (2004) reported the geneticcorrelation between density and MOE in radiata pine rang-ing from 0.44 to 0.64. Since MOE and density (and MOEand MFA) are favourably correlated and both density andMOE are under genetic control in the current study, then se-lection for MOE directly as opposed to selecting for itscomponent traits (density and MFA) should yield improve-ments in stiffness of juvenile radiata pine, as well as densityand MFA, as previously suggested by Dungey et al. (2006).
The genetic and phenotypic correlations between the same
wood quality trait measured in different rings or the whole-core measurement were also determined (Tables 6–8). Ge-netic correlations were always greater than phenotypic cor-relations for all traits throughout the core profile. Generally,the highest correlations for the same trait measured in differ-ent rings were observed between measurements in adjacentrings and were lowest between rings furthest apart. All ofthe genetic correlations were positive and significant withlow standard errors. Whole-core measurements of density,MFA, and MOE were highly correlated with individual ringmeasurements. For example, the genetic correlation betweenwhole-core area-weighted density and the mean density inthe pith was 0.84 and was >0.90 throughout the rest of thecore profile (Table 6). The genetic correlation betweenwhole-core MFA and individual-ring measurements rangedfrom 0.62 (pith) to 0.97 (ring 3) (Table 7). Similarly, the ge-netic correlation between whole-core area-weighted MOEand MOE in the rings reached 0.95 by ring 2 (Table 8).This is highly relevant for selection and breeding in that im-provement in juvenile wood properties across the entire pro-file of the corewood, including the innermost rings, can beachieved by selecting for a mean juvenile whole-core meas-urement.
ConclusionThere are two main approaches available for improving
juvenile wood quality in radiata pine. First, a reduction inthe proportion of juvenile wood in the tree would improveoverall wood quality since mature wood, or outerwood, hasmore desirable properties (Gapare et al. 2006). Second, im-
Table 7. Genetic (above diagonal) and phenotypic (below diagonal) correlations among radiata pine individual ring MFA mea-surements and the area-weighted MFA of the whole core from trials BR9611 and BR9705 (joint-site analyses).
Pith Ring 1 Ring 2 Ring 3 Ring 4 Ring 5 Ring 6* Whole core
Pith 0.8 (0.07) 0.64 (0.10) 0.52 (0.11) 0.46 (0.12) 0.46 (0.12) 0.65 (0.11) 0.62 (0.10)Ring 1 0.58 (0.02) 0.91 (0.03) 0.76 (0.06) 0.70 (0.07) 0.62 (0.09) 0.67 (0.09) 0.83 (0.05)Ring 2 0.41 (0.02) 0.75 (0.03) 0.94 (0.02) 0.90 (0.03) 0.85 (0.05) 0.90 (0.04) 0.96 (0.02)Ring 3 0.35 (0.03) 0.62 (0.02) 0.83 (0.01) 0.98 (0.01) 0.96 (0.02) 0.97 (0.02) 0.97 (0.01)Ring 4 0.34 (0.03) 0.54 (0.02) 0.7 (0.01) 0.89 (0.01) 0.99 (0.01) 0.98 (0.01) 0.96 (0.01)Ring 5 0.33 (0.03) 0.49 (0.02) 0.65 (0.02) 0.81 (0.01) 0.91 (0.01) 0.97 (0.01) 0.93 (0.02)Ring 6 0.34 (0.03) 0.50 (0.03) 0.65 (0.02) 0.77 (0.02) 0.83 (0.01) 0.92 (0.01) 0.98 (0.01)Whole core 0.47 (0.02) 0.75 (0.01) 0.87 (0.01) 0.91 (0.01) 0.88 (0.01) 0.85 (0.01) 0.86 (0.01)
Note: Standard errors are given in parentheses. MFA, microfibril angle.*Estimates based on analyses involving trial BR9611 only.
Table 8. Genetic (above diagonal) and phenotypic (below diagonal) correlations among radiata pine individual ring MOE mea-surements and the area-weighted MOE of the whole core from trials BR9611 and BR9705 (joint-site analyses).
Pith Ring 1 Ring 2 Ring 3 Ring 4 Ring 5 Ring 6* Whole core
Pith 0.83 (0.06) 0.65 (0.09) 0.54 (0.10) 0.52 (0.10) 0.52 (0.10) 0.63 (0.11) 0.65 (0.08)Ring 1 0.64 (0.02) 0.87 (0.04) 0.75 (0.06) 0.67 (0.08) 0.62 (0.09) 0.74 (0.08) 0.82 (0.05)Ring 2 0.43 (0.02) 0.76 (0.01) 0.94 (0.02) 0.89 (0.03) 0.83 (0.05) 0.91 (0.04) 0.95 (0.02)Ring 3 0.35 (0.03) 0.62 (0.02) 0.82 (0.01) 0.96 (0.02) 0.92 (0.03) 0.98 (0.02) 0.97 (0.01)Ring 4 0.31 (0.03) 0.53 (0.02) 0.70 (0.01) 0.83 (0.01) 0.97 (0.01) 0.99 (0.01) 0.95 (0.02)Ring 5 0.29 (0.03) 0.49 (0.02) 0.64 (0.02) 0.74 (0.01) 0.85 (0.01) 0.97 (0.02) 0.92 (0.02)Ring 6* 0.30 (0.03) 0.49 (0.03) 0.62 (0.02) 0.70 (0.02) 0.78 (0.01) 0.80 (0.01) 0.99 (0.01)Whole core 0.44 (0.02) 0.71 (0.02) 0.83 (0.01) 0.89 (0.01) 0.90 (0.01) 0.85 (0.01) 0.83 (0.01)
Note: Standard errors are given in parentheses. MOE, modulus of elasticity.*Estimates based on analyses involving trial BR9611 only.
2172 Can. J. For. Res. Vol. 37, 2007
# 2007 NRC Canada
provements in wood quality of the juvenile core, such as in-creased density and stiffness and reduced MFA, will lead toimprovements in the overall quality of the wood. The treesused in the current study were 8 to 9 years old, and conse-quently, the whole-core samples were entirely juvenilewood. Therefore, this study was tailored to address improve-ment in the quality of juvenile wood properties for Austral-ian radiata pine breeding programs.
All of the juvenile wood properties, density, MFA, andMOE, were found to be heritable in the radiata pine growingat two locations in Australia. These results, in conjunctionwith low genotype by environment interaction for juvenilewood properties, make these traits quite amenable to selec-tion, especially if these trends hold for other regions andsites. Favourable genetic correlations between MOE andMFA and MOE and density imply that improvement of mul-tiple juvenile wood properties is possible. For instance, se-lection for MOE directly will result in an improvement inoverall stiffness of the juvenile wood in radiata pine, withthe consequences of an increase in density and a reductionin MFA. Conversely, density is the wood quality trait thatis most likely to be incorporated in tree improvement pro-grams because it is relatively easy and inexpensive to meas-ure compared with other wood properties. Because densityand MOE are positively correlated, selection for densityshould result in genetic improvement in the overall stiffnessof juvenile wood in radiata pine. Furthermore, genetic corre-lations between juvenile wood properties measured in indi-vidual rings and the whole-core measurement weresignificant and positive, indicating that selection based on amean whole-core value of a juvenile wood trait, density forexample, will result in improvement of density in individualrings throughout the core including the innermost rings.
Although favourable genetic correlations were observedbetween juvenile wood quality traits, all of the wood proper-ties measured were unfavourably correlated with growth. Be-cause of unfavourable genetic correlations between juvenilewood quality traits and growth rate, selection for increaseddensity and stiffness or reduced MFA in the absence of se-lection for growth will result in a reduction (genetic loss) indiameter growth. Breeders and forest managers will have tostrike a balance between overall wood quality and growth,and geneticists should develop breeding strategies to dealwith such negative genetic correlations. Slight reductions ingrowth may be of little consequence when considering thegenetic gains in juvenile wood properties. There may be op-portunities for selection of correlation breakers for bothbreeding and deployment, thus improving juvenile woodproperties without adversely affecting growth in radiatapine. Developing breeding objectives may be the first stepin dealing with these negative genetic correlations. If wecan further understand the genetic basis of these negativegenetic correlations using molecular tools (e.g., identifyingpleiotropic alleles or genes associated with antagonisticallycorrelated traits), more efficient breeding strategies may bedeveloped to circumvent these antagonistic genetic correla-tions by crossing genotypes with a desirable suite of al-leles.
AcknowledgementsThis work forms part of the Juvenile Wood Initiative, a
collaborative project among Ensis/CSIRO, the SouthernTree Breeding Association, the Forest and Wood ProductsResearch and Development Corporation, the QueenslandDepartment of Primary Industries-Forestry, and ArborGen.The authors express sincere appreciation to Hancock Victo-ria Plantation Pty Ltd. and Green Triangle Timber ProductsLtd. for maintaining two genetic trials used for the study andfor assistance with sampling.
ReferencesBannister, M.H., and Vine, M.H. 1981. An early progeny trial in
Pinus radiata. 4. Wood density. N. Z. J. For. Sci. 11: 221–243.Burdon, R.D., and Low, C.B. 1992. Genetic survey of Pinus ra-
diata. 6: Wood properties: variation, heritability, and interrela-tionships with other traits. N. Z. J. For. Sci. 22: 228–245.
Burdon, R.D., Kibblewhite, R.P., Walker, J.C.F., Megraw, R.A.,Evans, R., and Cown, D.J. 2004. Juvenile versus mature wood: anew concept, orthoganal to corewood versus outerwood, withspecial reference to Pinus radiata and P. taeda. For. Sci. 50:399–415.
Cotterill, P.P., and Dean, C.A. 1990. Successful tree breeding withindex selection. Division of Forestry and Forest Products, Com-monwealth Scientific and Industrial Research Organisation(CSIRO), Collingwood, Victoria, Australia. p. 80.
Cown, D.J. 1992. Corewood (juvenile wood) in Pinus radiata —Should we be concerned? N. Z. J. For. Sci. 22: 87–95.
Cown, D., and van Wyk, L. 2004. Profitable wood processing —What does it require? Good wood! N. Z. J. For. 49: 10–15.
Cown, D.J., Young, G.D., and Burdon, R.D. 1992. Variation inwood characteristics of 20-year-old half-sib families of Pinusradiata. N. Z. J. Sci. 22: 63–76.
Dean, C.A. 1990. Genetics of growth and wood density in radiatapine. Ph.D. dissertation, The University of Queensland, Bris-bane, Australia. p. 78.
Dean, C.A., Cotterill, P.P., and Cameron, J.N. 1983. Genetic para-meters and gains expected from multiple trait selection of ra-diata pine in eastern Victoria. Aust. For. Res. 13: 271–278.
Donaldson, L.A. 1997. Microfibril angle: its influence on woodproperties, and prospects for improvement in radiata pine. InProceedings of the Wood Quality Workshop 1995. New ZealandForest Research Institute, Rotorua, New Zealand. FRI Bull. 201.pp. 20–25.
Donaldson, L.A., and Burdon, R.D. 1995. Clonal variation and re-peatability of microfibril angle in Pinus radiata. N. Z. J. For.Sci. 25: 164–174.
Dungey, H.S., Matheson, A.C., Kain, D., and Evans, R. 2006. Ge-netics of wood stiffness and its component traits in Pinus ra-diata. Can. J. For. Res. 36: 1165–1178. doi:10.1139/X06-014.
Evans, R. 1994. Rapid measurement of the transverse dimensionsof tracheids in radial wood specimens from Pinus radiata. Holz-forschung, 48: 168–172.
Evans, R. 2006. Wood stiffness by X-ray diffractometry. In Char-acterization of the cellulosic cell wall. Edited by D. Stokke andL. Groom. Blackwell Publishing. pp. 138–148.
Evans, R., Ilic, J., and Matheson, A.C. 2000. Rapid estimation ofsolid wood stiffness using SilviScan. In Proceedings of the 26thForest Products Research Conference, 19–21 June 2000, CSIROForestry and Forest Products, Clayton, Australia. pp. 49–50.
Gapare, W.J., Wu, H.X., and Abarquez, A. 2006. Genetic control inthe time of transition from juvenile wood to mature wood in Pi-nus radiata D. Don. Ann. For. Sci. 63: 871–878. doi:10.1051/forest:2006070.
Gilmour, A.R., Gogel, B.J., Cullis, B.R., and Thompson, R. 2005.
Baltunis et al. 2173
# 2007 NRC Canada
ASReml User Guide. Release 2.0. VSN International Ltd., He-mel Hempstead, UK. p. 267.
Hannrup, B., Cahalan, C., Chantre, G., Grabner, M., Karlsson, B.,Le Bayon, I., Jones, G.L., Muller, U., Pereira, H., Rodrigues,J.C., Rosner, S., Rozenberg, P., Wilhelmsson, L., and Wimmer,R. 2004. Genetic parameters of growth and wood quality traitsin Picea abies. Scand. J. For. Res. 19: 14–29. doi:10.1080/02827580310019536.
Herman, M., Dutilleul, P., and Avella-Shaw, T. 1999. Growth rateeffects on intra-ring and inter-ring trajectories of microfibril an-gle in Norway spruce (Picea abies). IAWA J. 20: 3–21.
Jayawickrama, K.J.S. 2001. Genetic parameter estimates for radiatapine in New Zealand and New South Wales: a synthesis of re-sults. Silvae Genet. 50: 45–53.
Jayawickrama, K.J.S., and Carson, M.J. 2000. A breeding strategyfor the New Zealand radiata pine breeding cooperative. SilvaeGenet. 49: 82–90.
Kendall, M.G., and Stuart, A. 1963. The advanced theory of statis-tics. Vol. 1. Hafner, New York. p. 433.
Kibblewhite, R.P., and Lloyd, J.A. 1983. Interrelations between thewood, fibre, and kraft pulp properties of new-crop radiata pine12-year-old thinnings. FRI Bull. For. Res. Inst. N.Z. For. Serv.Bull. 44: 19.
Kretschmann, D.E., and Bendtsen, B.A. 1992. Ultimate tensilestrength and modulus of elasticity of fast-grown plantation lo-blolly pine lumber. Wood Fiber Sci. 24: 189–203.
Kumar, S. 2004. Genetic parameter estimates for wood stiffness,strength, internal checking, and resin bleeding for radiata pine.Can. J. For. Res. 34: 2601–2610. doi:10.1139/x04-128.
Kumar, S., Jayawickrama, K.J.S., Lee, J., and Lausberg, M. 2002.Direct and indirect measures of stiffness and strength show highheritability in a wind-pollinated radiata pine progeny test inNew Zealand. Silvae Genet. 51: 256–261.
Li, L., and Wu, H.X. 2005. Efficiency of early selection for rota-tion-aged growth and wood density traits in Pinus radiata. Can.J. For. Res. 35: 2019–2029. doi:10.1139/x05-134.
Lindstrom, H., Harris, P., Sorensson, C.T., and Evans, R. 2004.Stiffness and wood variation of 3-year old Pinus radiata clones.Wood Sci. Technol. 38: 579–597. doi:10.1007/s00226-004-0249-1.
Lundgren, C. 2004. Microfibril angle and density patterns of ferti-lized and irrigated Norway spruce. Silva Fenn. 38: 107–117.
Matheson, A.C., Eldridge, K.G., Brown, A.G., and Spencer, D.J.1986. Wood volume gains from first-generation radiata pineseed orchards. Forest Research, Commonwealth Scientific andIndustrial Research Organisation (CSIRO), Canberra, Australia.DFR User Series No. 4. p. 13.
Megraw, R.A. 1985. Wood quality factors in loblolly pine. The in-fluence of tree age, position in tree, and cultural practise onwood specific gravity, fiber length, and fibril angle. TAPPIPress, Atlanta, Georgia.
Megraw, R.A., Leaf, G., and Bremer, D. 1998. Longitudinal shrink-age and microfibril angle in loblolly pine. In Microfibril anglein wood. Edited by B.G. Butterfield. University of CanterburyPress, Christchurch, New Zealand. pp. 27–61.
Megraw, R.A., Bremer, D., Leaf, G., and Roers, J. 1999. Stiffnessin loblolly pine as a function of ring position and height, and itsrelationship to microfibril angle and specific gravity. In Connec-tion between silviculture and wood quality through modellingapproaches, IUFRO W.P.S5.01-04, 5–12 August 1999, LaLonde-Les Maures, France. Edited by G. Nepveu. INRA Edi-tons, Versailles, France. pp. 341–349.
Myszewski, J.H., Bridgewater, F.E., Lowe, W.J., Byram, T.D., andMegraw, R.A. 2004. Genetic variation in the microfibril angle ofloblolly pine from two test sites. South J. Appl. For. 28: 196–204.
Namkoong, G. 1979. Introduction to quantitative genetics in for-estry. USDA For. Serv. Tech. Bull. No. 1588. p. 342.
Powell, M.B., McRae, T.A., Wu, H.X., Dutkowski, G.W., and Pil-beam, D.J. 2004. Breeding strategy for Pinus radiata in Austra-lia. In Forest Genetics and Tree Breeding in the Age ofGenomics: Progress and Future, Proceedings of the 2004 IUFROJoint Conference of Division 2, 1–5 November 2004, Charles-ton, South Carolina [online]. Edited by B. Li and S. McKeand.Available from http://www.ncsu.edu/feop/iufro_genetics2004/proceedings.pdf .
Sultech Report. 1999. Benefits from CSIRO Research for the for-estry, wood and paper industries sector: impact analysis andevaluation. Commonwealth Scientific and Industrial ResearchOrganisation (CSIRO), Canberra, Australia. p. 82.
Wright, P.J., and Eldridge, K.G. 1985. Profitability of using seedfrom the Tallaganda radiata pine seed orchard. Appita J. 38:341–344.
Wu, H.X., and Matheson, A.C. 2005. Genotype by environment in-teractions in an Australia-wide radiata pine diallel mating ex-periment: implications for regionalized breeding. For. Sci. 51:29–40.
Wu, H.X., Yang, J., McRae, T.A., Li, L., and Powell, M.B. 2004.Genetic relationship between breeding objective and early selec-tion criterion traits in Australia radiata pine population. Com-monwealth Scientific and Industrial Research Organisation(CSIRO), Canberra, Australia. CSIRO For. For. Tech. Rep.1402 and Southern Tree Breeding Association Tech. Rep.TR04–01.
Wu, H.X., Powell, M.B., Yang, J.L., Ivkovic, M., and McRae, T.A.2007. Efficiency of early selection for rotation-aged wood qual-ity traits in radiata pine. Ann. For. Sci. 64: 1–9. doi:10.1051/forest:2006082.
Zobel, B.J. 1981. Wood quality from fast grown plantations.TAPPI J. 64: 71–74.
Zobel, B.J., and van Buijtenen, J.P. 1989. Wood variation: it’scauses and control. Springer-Verlag, Berlin. p. 363.
2174 Can. J. For. Res. Vol. 37, 2007
# 2007 NRC Canada
