Ann. For. Sci. 63 (2006) 813–821
c ? INRA, EDP Sciences, 2006
Genetic parameters for lignin, extractives and decay
in Eucalyptus globulus
Fiona S. Pa*, Brad M. Pa, René E. Va, Carolyn A. Rb
aSchool of Plant Science and Cooperative Research Centre for Forestry, University of Tasmania, Private Bag 55, Hobart 7001, Tasmania, Australia
bForests NSW, PO Box 46, Tumut, NSW, Australia
(Received 8 August 2005; accepted 1 February 2006)
Abstract – Eucalyptus globulus is grown in temperate regions of the world for pulp production. The chemical and physical properties of its wood make
it highly suited to this purpose. This study analysed genetic variation in lignin content, extractives content and decay, for nine localities of E. globulus.
Heritability estimates were also obtained, and the relationships between these traits and physical wood traits and growth were examined. Significant
genetic variation was found between localities for lignin content (Klason lignin and acid-soluble lignin contents) and decay. The only trait for which
significant variation between families within locality was detected was acid-soluble lignin content, which resulted in this trait also having the highest
narrow-sense heritability (0.51 ± 0.26). However, family means heritabilities were high for lignin content, extractives content and decay (0.42–0.64).
The chemical wood traits were strongly correlated with each other both phenotypically and genetically, with important correlations found with density
and microfibril angle. Correlations suggested that during selection for the breeding objective traits, it is likely that favourable states in the chemical
wood traits, decay resistance and fibre properties are concurrently being selected, whereas growth may be selected for independently. This initial study
provides a stepping stone for future studies where particular localities of the breeding population may be characterised further.
correlation / eucalypt / genetic variation / heritability / lignin
Résumé – Paramètres génétiques pour la lignine, les extractibles et la pourriture chez Eucalyptus globulus. Eucalyptus globulus est cultivé dans
de nombreuses régions tempérées pour la production de pâte à papier. Les propriétés physiques et chimiques de ce bois en font un matériau très apprécié
pour cetteutilisation.Cetteétude analyse les variations génétiques de lateneur enlignine, encomposés extractibles et de lasensibilitéà ladécomposition
du bois de neuf provenances de Eucalyptus globulus. L’héritabilité de ces propriétés ainsi que leurs relations avec les caractéristiques physiques du bois
et la croissance ont été examinées. Des différences inter-provenances significatives ont été détectées pour les teneurs en lignine (lignines dosées par la
méthode de Klason, ou lignines soluble en solution acide) et la vitesse de décomposition. Le seul caractère qui a présenté une variation significative
entre familles dans les provenances a été la teneur en lignines solubles en solution acide. De ce fait, une forte héritabilité (au sens strict) a été détectée
pour ce caractère (0,51±0,26). Cependant, les héritabilitésmoyennes dans les familles étaient élevées pour lateneur en lignine, les teneurs en composés
extractibles et la vitesse de décomposition (0,42–0,64). Les caractéristiques chimiques du bois étaient fortement inter-corrélées au niveau phénotypique
et génétique, avec des corrélations importantes également avec la densité et l’angle des microfibrilles. Les corrélations suggèrent que durant la sélection
de caractères objectifs d’amélioration, des traits favorables associant caractéristiques chimiques du bois, résistance à la décomposition et propriétés des
fibres puissent être sélectionnés simultanément, alors que la croissance doit faire l’objet d’une sélection indépendante. Cette étude constitue une base
pour de futurs travaux permettant une caractérisation plus fine de provenances particulières dans cette population de sélection.
corrélation / eucalyptus / variation génétique / héritabilité / lignine
Eucalyptus globulus is grown for pulp production in tem-
perate Australia and other parts of the world, including South
America, southern Europe, Africa and Asia. Considerable ge-
netic variation has been identified between the subraces of E.
globulus for a wide range of traits, including growth and both
physical and chemical wood properties [1,10,23,24].Some of
this variation has been exploited in breeding programs for the
selection of superior trees. When selecting trees for pulp pro-
duction, only a few traits are currently examined in Australia,
with the focus on increases in volume per hectare, basic den-
sity and pulp yield [1, 16, 17]. Although selection for these
traits gives an increase in the pulp yield per hectare, many
*Corresponding author: firstname.lastname@example.org
other physical and chemical wood properties are important to
kraft pulping, and variations in these can be conduciveto min-
imising the costs or environmental impacts of the process.
Kraft pulping generally involves the removal of most of
the extractives, approximately 80% of the lignin and approxi-
mately 50% of the hemicellulose from the cellulose fibres us-
ing alkali . For the production of high quality paper, the
pulp is further bleached to remove the residual lignin, which
is responsible for turning the paper yellow through oxidation
and light absorption . The lignin and extractives contents
of wood, are traits that are fast being recognised as having im-
portancein minimisingthe costs andenvironmentalimpacts of
kraft pulping. As lignin and extractives are the primary waste
products of the pulping process, lower levels in the wood will
result in faster delignification and/or a reduction in the use of
Article published by EDP Sciences and available at http://www.edpsciences.org/forest or http://dx.doi.org/10.1051/forest:2006080
814F.S. Poke et al.
chemicals and energy. This will help minimise the production
of pollutants from the pulping process.
Studies into the genetic variation and heritability of lignin
and extractives have been limited in Eucalyptus until recently.
This was mainly due to the expensive and time-consumingna-
ture of the chemical assay used to measure these traits [2,3].
More recently, simple and cost-effective techniques have been
developed for predicting these traits, involving near infrared
reflectance (NIR) analysis on ground wood cores [4,30]. This
has been found to be an effective technique for reliably pre-
dicting these traits in large numbers of samples. A small num-
ber of studies have suggested that there is genetic variation in
lignin and extractives contents in E. globulus, although these
involvedonly three or four provenancesand a small numberof
individuals [24,41]. Studies investigating the correlations be-
tween thechemicalwoodtraits andotherwoodpropertiesinE.
globulus, have also been limited by small sample sizes and to
small numbers of traits, and have involved phenotypiccorrela-
tions only[23,39].Geneticcorrelationshavenot beenreported
for these traits in E. globulus. To fully explore the scope of
variation within the chemical traits and their genetic control,
large numbers of individuals, families and provenancesare re-
quired which encompass the range of E. globulus.
The susceptibility of trees to wood decay is important as it
may impact on pulp yield in two ways: firstly, the plant de-
fence response may lead to an increase in the amount of lignin
andextractivespresentin thewood,whichwill reducethe pulp
yield; secondly, decay leads to degradation of the wood caus-
ing a breakdown of the cellular structure . This decay can
be caused by pathogen infection of exposed, vulnerable tis-
sue following wounding, or through attack of the heartwood
(heart rot), which is incapable of an active response due to the
lack of living cells . The impact of decay will also depend
upon whether the decay organisms are feeding on cellulose or
lignin. Two types of fungi generally are responsible for decay,
brown rot fungi which degrade cellulose and white rot fungi
which degrade lignin. Decay can be observed in “pockets”
that are compartmentalised by a reaction zone (purple/pink
coloured boundary between healthy and decayed wood) and
discolouration of the surrounding wood, or as rotting of the
heartwood [29,40]. Fungal attack has been found to be asso-
ciated with increases in lignin due to its resistance to degrada-
tion by pathogens, in E. gunnii , and increases in extrac-
tives which contain antimicrobial compounds, in E. nitens .
In E. globulus genetic variation in decay susceptibility and its
relationship to other wood properties has not been examined.
Improving the chemical wood properties of tree species
through breeding requires genetic variation to be present for
selection. It is also necessary to understand how the traits to
be improved are related to one another and to other traits that
are not currently being selected, so that when an increase in
one trait is selected for,the potential effects on othertraits may
be predicted. A study conducted by Apiolaza et al.  exam-
ined the variation in growth and wood traits as well as their
correlations using 188 individuals of E. globulus from 35 fam-
ilies and eight subraces, which are currently part of the single
breeding population of the Australian national breeding pro-
gram. The traits examined included diameter at breast height
over bark (DBH), basic density (BD), mean fibre length (FL),
mean microfibril angle (MFA) which is the average angle of
the cellulose microfibrillar helix relative to the longitudinal fi-
bre axis , predicted pulp yield (PY) and cellulose content
(CELL). The current study aimed to build on that of Apiolaza
et al.  with a particular focus on the chemical wood prop-
erties. Using the same open-pollinated progenies grown in a
field trial, we examined the variation in and the heritability of
lignin content (LIG), extractives content (EXTR) and extent
of decay (DEC) between and within nine localities of E. glob-
ulus originating from around Tasmania and south-east Victo-
ria. Phenotypic and genetic correlations were also determined
amongst these traits and with the growth and wood traits of
Apiolaza et al. . The relationship between these chemical
wood traits and with the physical wood traits, wood decay and
growth has not been examined before in E. globulus and will
providean indicationof how multiple traits are affected during
the selection of superior trees.
2. MATERIALS AND METHODS
2.1. Plant material
Wood samples of Eucalyptus globulus were collected from a base
population field trial located at West Ridgley, Tasmania (Gunns Ltd).
This trial was established in 1989 based on the CSIRO Australian
Tree Seed Centre collection and is comprised of open-pollinated fam-
ilies [12,13]. The trial was an incomplete block design with 451 fam-
ilies in five replicates, each with 17 incomplete blocks, and two-tree
row plots . A total of 177 trees from 37 families (Tab. I) were sam-
pled to cover the same range of eight subraces sampled by Apiolaza
et al. , with one tree or occasionally two trees per plot sampled.
Due to the fact that only a subset of families was sampled the trial
was treated as a randomised complete block design for analysis. The
locality denoted North-east Tasmania comprises two localities, Royal
Georgeand German Town, which weremerged because of smallsam-
ple sizes and their close proximity. Two bark-to-bark wood cores
were taken from each tree approximately 10 cm above the previous
core sites taken by Apiolaza et al.  two years before, according to
the method described by Raymond et al. .
2.2. Wood and growth measurements
Measurements for BD, MFA, FL, PY and CELL already existed
for these trees at age 11 years . Additional measurements were
taken for BD and DBH, and measurements were obtained for DEC,
LIG and EXTR all at age 13 years.
DEC was recorded for each core as the percentage of the core
with heart rot, pocket decay and/or discolouration and results were
averaged for the tree. For statistical analysis the different types of
decay data were grouped, with 0 indicating no decay followed by
10% intervals thereafter, and class midpoints were used for analysis.
Due to the presence of decay in the pith for many of the cores, partial
cores (the outer quarters of each core free of decay) were used for
further BD, LIG and EXTR measurements.
BD was determined for one core from each tree using the water
displacement method , by submerging the partial cores in cold
water for approximately two days, removal of remaining bark and
Lignin, extractives and decay in E. globulus
Table I. Breakdown of subraces (as classified by Dutkowski and Potts ), localities and families of E. globulus used in this study from the
base population trial at West Ridgley, Tasmania.
Flinders Island, Tasmania
King Island, Tasmania
Strzelecki Foothills, Victoria
Strzelecki Ranges, Victoria
Western Otways, Victoria
Locality Number of families
Number of individuals
Central Flinders Island
South King Island
North Maria Island
excess water followed by volume (V) measurements. The mass (M)
of each core was taken after drying at 105◦C for approximately two
days. BD was calculated using the following formula:
Theremaining partial coreswereused todevelop theNIR calibrations
reported in Poke et al.  for total lignin (TLIG), acid-soluble lignin
(ASLIG) and Klason lignin (KLIG) contents (TLIG = ASLIG +
KLIG) plus EXTR. These calibrations were based on chemical mea-
surements for 54 to 61 samples and had good correlation coefficients
(0.62–0.93), and predicted and laboratory values for the validation set
of samples were highly correlated (0.83–0.99) . The calibrations
were used to predict these traits for the remainder of the individuals
in the data set.
2.3. Statistical analysis
Variance components for BD, DBH, ASLIG, KLIG, TLIG, EXTR
and DEC were estimated using the MIXED procedure in SAS (Ver-
sion 9.1, SAS Institute Inc.), with locality fitted as a fixed effect, and
family withinlocalities, replicateand residual withinlocalities as ran-
dom effects. Locality least square means and the differences between
them were also calculated using the MIXED procedure in SAS, with
a Tukey-Kramer adjustment applied for multiple comparisons.
The individual narrow-sense (h2
tabilities of BD, DBH, ASLIG, KLIG, TLIG, EXTR and DEC were
estimated using ASREML , with the fixed locality term removed
from the model in the latter case. h2
itability within localities which is used operationally to predict ge-
netic gains from within locality selection. H2
heritability which indicates the gain that would be made from select-
ing the best families across all localities for deployment. h2
were estimated as :
op) and family mean (H2
oprefers to the narrow-sense her-
fmis the family means
component estimated assuming a coefficient of relatedness within
open-pollinated families of 0.4, after first adjusting the additive re-
lationship matrix for a 30% selfing rate ;
k = harmonic mean number of trees per family.
Trait correlations were determined amongst the age 13 measure-
ments of BD, DBH, ASLIG, KLIG, TLIG, EXTR and DEC and with
the traitsof Apiolaza et al. . Phenotypic correlations (Pearsons cor-
relation matrix) amongst individuals were determined in SAS using
the CORR procedure. Additive genetic correlations could not be es-
timated using ASREML  directly, as bivariate models failed to
converge due to the small sample size. However, as an approxima-
tion of the genetic correlations, Pearsons correlation matrices were
obtained using family means adjusted for locality differences and for
the nine locality means using the CORR procedure in SAS.
add(loc)= additive genetic variation within locality variance
f= family variance component calculated across localities;
e= residual variance component;
3.1. Trait statistics and variances
The number of individuals measured for each trait and the
statistics for each trait are detailed in Table II and include the
subset of measurements from Apiolaza et al. . Sixty-nine
percent of samples were found to have decay symptoms. Vari-
ation in the traits measured in this study, between replicates,
localities and family within localities, are detailed in Table III.
No significant variation was detected at any level for BD,
TLIG and EXTR (Tab. III). Locality was a significant source
of variation for DEC, DBH, KLIG and ASLIG (Tab. III). Of
these four traits, only two had significant differences between
the locality least square means following Tukey-Kramer ad-
justment (Tab. IV). For DEC, South King Island had signif-
icantly more decay than five other localities including Bow-
den Road (P < 0.001), Madalya Road (P < 0.002), Central
Flinders Island (P < 0.003), North-east Tasmania (P < 0.004)
and Cannan Spur (P < 0.023). For ASLIG, South Geeve-
ston and South King Island were significantly different to each
other(P < 0.01).Significant variationbetweenfamilies within
locality was detected for ASLIG only.
816F.S. Poke et al.
Table II. Statistics for growth and wood measurements of individual trees for the E. globulus base population trial at West Ridgley, Tasmania.
Mean fibre length at age 11 (FL)
Mean microfibril angle at age 11 (MFA)
Predicted pulp yield at age 11 (PY)
Cellulose content at age 11 (CELL)
Basic density at age 11 (BD)
Basic density at age 13 (BD)
Diameter at breast height at age 13 (DBH)
Klason lignin content at age 13 (KLIG)
Acid-soluble lignin content at age 13 (ASLIG)
Total lignin content at age 13 (TLIG)
Extractives content at age 13 (EXTR)
Extent of decay at age 13 (DEC)
Table III. Analyses of variance for growth and wood traits at age 13 years between replicates, localities, and families within localities, plus
estimates of the heritability of within locality variation, and family means heritability, for these traits in the samples from the E. globulus base
population trial at West Ridgley, Tasmania. Probability values are denoted *** P < 0.001, * P < 0.05 and ns = non-significant.
Trait df Basic
Replicate Z value
Locality F value
Family [locality] Z
Z values are random terms and F values depict fixed terms.
aVariance component was at the boundary of the parameter space.
3.2. Heritability estimates
Narrow-sense heritability estimates had large standard er-
rors due to the lack of significant variation between fami-
lies within localities for most traits, no doubt reflecting the
small sample size (Tab. III). Moderately high heritability val-
ues wereobtainedforASLIG (0.51 ± 0.26)andEXTR(0.35 ±
0.23), although ASLIG was the only trait where significant
variationbetween families within localities was detected.Both
BD (0.24 ± 0.26) and TLIG (0.29 ± 0.23) showed moderate
heritabilities, with KLIG (0.13 ± 0.20) showing little heri-
tability. Within locality variation in DEC and DBH was non-
heritable. The heritabilities of family means integrated both
within and between locality variation, and were somewhat
higher than the narrow-sense heritabilities due to the inclusion
of locality effects in the differences between families. ASLIG
(0.64 ± 0.10), EXTR (0.48 ± 0.14), TLIG (0.50 ± 0.14) and
DEC (0.50 ± 0.14) showed high estimates. BD (0.42 ± 0.19)
and KLIG (0.42 ± 0.16) had moderately high estimates, and
DBH a moderate estimate (0.19 ± 0.19) (Tab. III).
3.3. Trait correlations
Strong correlations were identified between the wood and
growth traits at locality, family and individual (phenotypic)
Lignin, extractives and decay in E. globulus
Table IV. Locality least square means and standard errors (in parenthesis) for growth and wood traits at age 13 years for samples from the
E. globulus base population trial at West Ridgley, Tasmania.
Localities with common letters for the same trait are not significantly different at P < 0.05 following Tukey-Kramer adjustment for multiple compar-
levels (Tab. V). Correlations between locality means and fam-
ily means with locality differences removed, represented ge-
netic based correlations. As expected, a strong, positive re-
lationship was identified between TLIG and its components
(ASLIGandKLIG) atmost levels, althoughKLIGandASLIG
were not significantly correlated. EXTR was strongly corre-
lated with lignin content for individuals, but the correlations
Genetic correlations were observed between EXTR and both
KLIG (families and localities) and TLIG (localities). KLIG,
TLIG and EXTR all had significant, negative, phenotypic and
genetic correlations with both CELL and PY, although these
were sometimes not significant at the locality level. Lignin
content showed significant negative correlations with BD at
the family level supported at both ages 11 and 13 years. TLIG
was also positively correlated with MFA at the individual and
family level, with ASLIG and KLIG correlated with MFA at
the individual level only. TLIG and DBH showed a weak neg-
ative correlation at the locality level only.
DEC was highly negatively correlated with BD only at age
11 years for individuals and localities, but not for families.
Significant genetic variation has been reported for BD at age
11 at the subrace level . When the South King Island local-
ity (particularly susceptible to decay) was removed from the
analysis, DEC and BD (age 11) were no longer correlated at
the locality level, but a significant correlation still remained at
the individual level (r = −0.230, P < 0.017). DEC had weak,
positive phenotypic correlations with PY and CELL, but there
were no significant genetic relationships. DEC showed nega-
tive relationships with KLIG, TLIG and EXTR for localities,
which were still significant for KLIG (r = −0.714, P < 0.047)
and EXTR (r = −0.717, P < 0.045) when South King Island
was removedfromtheanalysis.DEChada positivecorrelation
with FL at the family level only.
4.1. Variation in and heritability of wood properties
Two of the four chemical wood traits examined had
significant variation at either the locality or family within lo-
calitylevel,indicatingthereis geneticvariationwithin E.glob-
ulus. Useful heritability estimates were also obtained for sev-
eral traits despite their relatively large standard errors due to
the small sample size. Both Klason lignin and acid-soluble
lignin contents showed significant variation among localities,
which suggested improvement could be made through local-
ity selection. Surprisingly no locality differences were found
for total lignin content, although this is consistent with the
study of Miranda and Pereira  using five trees of E. glob-
ulus from each of four provenances. Acid-soluble lignin con-
tent showed significant variation for families and the highest
818F.S. Poke et al.
Table V. Correlations (Pearsons correlation matrix) amongst growth and wood traits for the E. globulus base population trial at West Ridgley,
Tasmania. L = correlations amongst the nine locality means (df = 7), F = correlations amongst family means (adjusted for locality differences;
df = 25 to 27) and I = phenotypic correlations amongst individuals (df = 106 to 155). Significant probability values are denoted *** P < 0.001,
**P < 0.01, * P < 0.05.
Extent of decay
Mean fibre length (FL)
Predicted pulp yield
Diameter at breast
Total lignin content
estimated narrow-sense (0.51) and family means (0.64) her-
itabilities. The only published narrow-sense heritability esti-
mate for lignin traits in E. globulus is for total lignin content
which was estimated to be very low at 0.09 ± 0.21 . The
moderate narrow-sense heritability estimate for total lignin
content from the current study (0.29), together with a high
family means heritability (0.50), suggest that lignin may be
under stronger genetic control than previously thought. Sup-
porting this is an estimate for the clonal heritability of lignin
content in E. globulus of 0.83 from Gominho et al. .
Extractives content was found to have a moderate narrow-
sense heritability (0.35 ± 0.23), but no statistically signif-
icant differences were found between or within localities.
Significant provenance effects for extractives content have
been found previously in E. globulus [24, 41], suggesting
provenance selection could be used to improve this trait. The
Lignin, extractives and decay in E. globulus
lack of variation in the current study may be due to different
provenances being used, or may be attributed to possible site
by genotype interactions affecting this trait. Kube  found
strong genotype by site interactions for extractives among 434
E. nitens trees from 40 families grown over three sites, with
heritability estimates found to vary between sites from low to
very high, suggesting that the factors causing extractives pro-
duction in some genotypes are very site specific. Miranda and
Pereira  found no site effects for extractives while the cur-
rent study found no replicate effects in E. globulus.
The heritability of and the variation in basic density and di-
ameter forE. globulushas been examinedextensivelyand
so will not be discussed in detail here. Basic density (age 13)
did not havesignificant variationat the family or locality level,
although the trends in locality means (King Island low and the
Strzelecki localities high) were consistent with previous stud-
ies that have reported significant differences [1,10,26]. This
suggested that the small sample size and the use of only the
outer part of the core reduced the power of the current study
and therefore significance would generally be underestimated.
All of the trees used in this study had been cored previ-
ously which meant tissue had potential exposure to infection
by wood decaying organisms. The West Ridgley site is also
a wet site which has been found to be a factor leading to an
increase in the incidence of decay . Localities differed
significantly in the extent of decay with South King Island
of genetic variation for decay resistance in E. globulus. The
two main races of E. globulus that have been used for plan-
tation growth in Australia, Strzelecki and King Island ,
were placed at either end of the range in decay as they have
been previously for basic density . The fast growing but
low density King Island trees were originally grown for pulp
production, however, Strzelecki and Western Otways became
preferred . It appears that high basic density trees now
selected in the breeding program may be more resistant to
decay. Although the narrow-sense heritability for decay was
estimated here as zero, a high family means heritability was
obtained (0.50 ± 0.14). Narrow-sense heritability estimates in
E. nitens have been found to vary between studies from 0.13
to 0.41 [20,42], and also between sites in a single study rang-
ing from 0.04 to 0.63 . The successful exclusion of decay
is likely the result of a number of traits including lignin and
extractives contents, and therefore environmental and site in-
fluences are likely to be strong .
4.2. Correlations amongst wood properties
Phenotypic correlations indicate the presence of relation-
ships between traits that may be due to a similar response to
environmental conditions or to genetic associations. Genetic
correlations are important for determining the potential for
concurrent or independent selection of traits. Correlations be-
tween family means (adjusted for locality differences) and be-
tween locality means, were used to give an indication of the
genetic associations for this dataset. No study has yet identi-
fied the genetic correlations among the chemical wood traits
(excluding pulp yield and cellulose content) and their corre-
lated effects on the physical wood traits and growth in E. glob-
Correlations amongst the chemical wood traits were of-
ten strong and as expected in terms of kraft pulping prop-
erties . A high pulp yield and cellulose content was as-
sociated with low extractives, Klason lignin and total lignin
contents at both the phenotypic and genetic levels. This was
consistent with the phenotypic correlations reported by Wallis
et al.  for 11 individuals of E. globulus. Miranda and
Pereira  examined 37 provenances of E. globulus and re-
ported a similar correlation between pulp yield and extrac-
tives content at the provenance level, but not with total lignin
content. No significant correlations were identified between
acid-soluble lignin content and Klason lignin content, consis-
tent with the findings of Miranda and Pereira  who sug-
gested differences in the lignin composition may be responsi-
ble. Lignin and extractives contents were generally positively
correlated here and Ona et al.  found similar relation-
ships in a within-tree study of two E. globulus individuals.
In E. nitens Kube and Raymond  reported a very high
negative genetic correlation between extractives and cellulose
contents. These studies collectively suggest that selection for
increased pulp yield or cellulose content are likely to result
in a reduction in lignin and extractives contents, which are
favourable responses for a pulpwood breeding objective.
The correlated effect of lignin on wood density is interest-
ing as density is one of the main selection traits in the E. glob-
ulus breeding program. Basic density at ages 11 and 13 were
significantly positively correlated at most levels (r = 0.56,
P < 0.01 for families and r = 0.65, P < 0.0001 for indi-
viduals), and both were negatively correlated with lignin con-
tent at the family level. No other studies have looked at the
relationship between lignin and basic density for larger sam-
ple sizes in Eucalyptus. However, a negative genetic correla-
tion has also been found between density and lignin content
in Pinus pinaster . It is therefore likely that favourable
lignin profiles are being indirectly selected along with high
basic density. Similar to other studies in E. globulus [23,28]
no apparent relationship was found between basic density and
extractives content, although there are reports of positive asso-
ciations in both E. globulus  and E. nitens .
tween microfibril angle and lignin content which is consistent
with observations for coniferous wood . This relationship
is thoughtto be due to the distributionof the microfibrilsabout
their preferredorientation being large when the microfibril an-
gle is large, therefore creating an imperfect alignment with
more room for lignin deposition . These results suggest
that a reducedmicrofibrilangle (whichgivesthe fibre a greater
tensile strength and decreases its shrinkage ) may be asso-
ciated with improved lignin profiles for pulping.
Decay resistance is unlikely to become a major focus for
selection in breeding programs for pulpwood, however, it is
an important issue in the production of solid wood . Un-
derstanding the genetic relationships between decay and the
chemical and physical wood traits, as well as growth, is there-
fore important. When examining relationships between decay
820F.S. Poke et al.
incidence or extent with wood chemistry, it is important to
distinguish between the chemistry found for normal healthy
wood, and that found in diseased wood or in the reaction zone
between healthy and diseased wood. It has been reported pre-
viously that the extractives and lignin contents are elevated
in response to decay in eucalypts [5,18], with the extractives
content found to be six times greater in the reaction zone com-
pared to healthy sapwood . Only negative locality level cor-
relations were foundin the current study between the extent of
have been found between extent of decay and extractives con-
tent in E. nitens , however, a negative relationship has
been found in E. delegatensis . Together these studies in-
dicate that increases in extractives and lignin contents may
only occur for diseased wood or in the reaction zone (both of
which were removed in the current study), and the surround-
ing, healthy wood has normal extractives and lignin levels.
A negative correlation between the extent of decay and ba-
sic density (age 11) was observed at the locality level, which
seemed to be the result of one locality (South King Island) that
appeared to be particularly susceptible to decay and is known
for its low basic density . However, the extent of decay
showed significant phenotypic correlations with basic density
(age 11, negative), even with the South King Island locality
removed from the analysis. Similar negative correlations have
also been found in E. delegatensis and E. grandis [27,43]. It
has been proposed that lower density wood has wider cell lu-
mina, and therefore a larger surface area is exposed to the en-
zymes of decay micro-organisms, and also the water and air
content in the wood may be at a level that promotes fungal
growth . A positive genetic correlation between the extent
of decay and mean fibre length was also found, and may sup-
port this idea. The lack of a significant correlation between
the extent of decay and basic density at age 13, may be be-
cause the decayed area of the core was removed prior to basic
density measurements and only partial cores were used. The
age 11 measures of wood density may have been taken be-
fore the formation of the decay and may be more indicative of
wood susceptibility to decay.
The combination of the chemical wood properties with the
physical wood properties of Apiolaza et al. , allows a pri-
mary analysis of the genetic variation of the most important
traits associated with pulp production, and how they are cor-
related with one another. This is the first study incorporating
such a large number of traits for E. globulus, although the re-
sults must be treated with some caution due to the small sam-
ple size. The results indicate that when selecting for the cur-
rent breeding objective traits of high basic density and pulp
yield , other traits beneficial to the pulping process may
concurrently be selected, including low lignin and extractives
contents, and a high cellulose content, as well as improved fi-
bre properties. Selection for high basic density may also result
in increased resistance to decay. Growth may be selected for
independently of most of the chemical wood properties and
Acknowledgements: The authors would like to thank Gunns Ltd for
access to field trials, Leon Savage for assistance with field sampling,
and also the Australian Research Council for support.
 Apiolaza L.A., Raymond C.A., Yeo B.J., Genetic variation of phys-
ical and chemical wood properties of Eucalyptus globulus, Silvae
Genet. 54 (2005) 160–166.
 Appita Standard, Klason lignin in wood and pulp, Australian Pulp
and Paper Industry Technical Association, Appita Standard P11s,
 Appita Standard, Organic solvent extractives in wood, pulp and
paper, Australian Pulp and Paper Industry Technical Association,
Appita Standard 1301.012s, 1994.
 Baillères H., Davrieus F., Pichavant F.H., Near infrared analysis as
a tool for rapid screening of some major wood characteristics in a
eucalyptus breeding program, Ann. For. Sci. 59 (2002) 479–490.
 Barry K.M., Pearce R.B., Evans S.D., Hall L.D., Mohammed C.M.,
Initial defence responses in sapwood of Eucalyptus nitens (Maiden)
following wounding and fungal inoculation, Physiol. Mol. Plant
Pathol. 58 (2001) 63–72.
 Barry K.M., Pearce R.B., Mohammed C.L., Properties of reaction
zones associated with decay from pruning wounds in plantation-
grown Eucalyptus nitens, For. Pathol. 30 (2000) 233–245.
 Bootle K.R., Wood in Australia: types, properties, uses, McGraw-
Hill, Sydney, 1983.
 Cotterill P.P., Brolin A., Improving Eucalyptus wood, pulp
and paper quality by genetic selection, in: Proceedings of the
IUFRO Conference on Silviculture and Improvement of Eucalypts,
Salvador, Brazil, August 24–29, 1997, pp. 1–13.
 Dutkowski G.W., Gilmour A.R., Borralho N.M.G., Modification of
theadditive relationshipmatrixforopen pollinatedtrials,in: IUFRO
International Symposium on Developing the Eucalypt of the Future,
Valdivia, Chile, September 10–15 2001.
 Dutkowski G.W., Potts B.M., Geographic patterns of genetic varia-
tion in Eucalyptus globulus ssp. globulus and a revised racial clas-
sification, Aust. J. Bot. 47 (1999) 237–263.
 Evans R., Hughes M., Menz D., Microfibril angle variation by scan-
ning X-ray diffractometry, Appita J. 52 (1999) 363–367.
 Gardiner C.A., Crawford D.F., Seed collections of Eucalyptus
globulus subsp. globulus for tree improvement purposes, CSIRO
Division of Forest Research, Canberra, 1987.
 Gardiner C.A., Crawford D.F., Seed collections of Eucalyptus glob-
ulus subsp. globulus Labill. for tree improvement purposes, CSIRO
Division of Forestry and Forest Products, Canberra, 1988.
 Gilmour A.R., Cullis B.R., Welham S.J., Thompson R., ASREML
computer software, NSW Agriculture, Orange, 1999.
 Gominho J., Rodrigues J., Almeida M.H., Leal A., Cotterill P.P.,
Pereira H., Assessment of pulp yield and lignin content in a
first-generation clonal testing of Eucalyptus globulus in Portugal,
in: Proceedings of the IUFRO Conference on Silviculture and
Improvement of Eucalypts, Salvador, Brazil, August 24–29, 1997,
 Greaves B.L., Borralho N.M.G., The influence of basic density and
pulp yield on the cost of eucalypt kraft pulping: a theoretical model
for tree breeding, Appita J. 49 (1996) 90–93.
 Greaves B.L., BorralhoN.M.G.,Raymond C.A.,Breeding objective
for plantation eucalypts grown for production of kraft pulp, For. Sci.
43 (1997) 465–472.
 Hawkins S., Boudet A., “Defence lignin” and hydroxycinnamyl al-
cohol dehydrogenase activities in wounded Eucalyptus gunnii, For.
Pathol. 33 (2003) 91–104.
Lignin, extractives and decay in E. globulus Download full-text
 Johnson G.R., Burdon R.D., Family-site interaction in Pinus ra-
diata: Implications for progeny testing strategy and regionalised
breeding in New Zealand, Silvae Genet. 39 (1990) 55–62.
 Kube P.D., Genetic improvement of the wood properties of
Eucalyptus nitens: Breeding to improve solid wood and pulp prop-
erties, Ph.D. thesis, University of Tasmania, Hobart, Australia,
 Kube P.D., Raymond C.A., Genetic parameters for Eucalyptus
nitens solid wood traits and relationships with pulpwood traits, in:
IUFROInternational Symposium on Developing theEucalypt of the
Future, Valdivia, Chile, September 10–15, 2001.
 Lopez G.A., Potts B.M., Dutkowski G.W., Apiolaza L.A., Gelid
P.E., Genetic variation and inter-trait correlations in Eucalyptus
globulus base population trials in Argentina, For. Genet. 9 (2002)
 Miranda I., Pereira H., Provenance effect on wood chemical com-
position and pulp yield for Eucalyptus globulus Labill, Appita J. 54
 Miranda I., Pereira H., Variation of pulpwood quality with prove-
nances and site in Eucalyptus globulus, Ann. For. Sci. 59 (2002)
 Mohammed C., Barry K., Battaglia M., Beadle C., Eyles A.,
Mollon A., Pinkard E., Pruning-associated stem defects in plan-
tation E. nitens and E. globulus grown for sawlog and veneer in
Tasmania, Australia, in: The future of eucalypts for wood pro-
duction, Launceston, Tasmania, Australia, March 19–24, 2000,
 Muneri A., Raymond C.A., Genetic parameters and genotype-by-
environment interactions for basic density, pilodyn penetration and
stem diameter in Eucalyptus globulus, For. Genet. 7 (2000) 317–
 Nelson N.D., Heather W.A., Wood color, basic density, and decay
resistance in heartwood of fast-grown Eucalyptus grandis Hill ex.
Maiden, Holzforschung 26 (1972) 54–60.
 Ona T., Sonoda T., Ito K., Shibata M., Relations between vari-
ous extracted basic densities and wood chemical components in
Eucalyptus globulus, J. Wood Sci. 44 (1998) 165–168.
 Pearce R.B., Tansley review No. 87. Antimicrobial defences in the
wood of living trees, New Phytol. 132 (1996) 203–233.
 Poke F.S., Wright J.K., Raymond C.A., Predicting extractives and
lignin contents in Eucalyptus globulus using near infrared re-
flectance analysis, J. Wood Chem. Technol. 24 (2004) 55–67.
 Pot D., Chantre G., Rozenberg P., Rodrigues J.C., Jones G.L.,
Pereira H., Hannrup B., Cahalan C., Plomion C., Genetic control of
pulp and timber properties in maritime pine (Pinus pinaster Ait.),
Ann. For. Sci. 59 (2002) 563–575.
 Potts B.M., Dutkowski G.W., Jordan G.J., Vaillancourt R.E.,
Providing a population genetic framework for exploitation of
eucalypt genetic resources: The case of Eucalyptus globulus,
in: Proceedings of the Australian Plant Breeding Conference,
Adelaide, South Australia, April 19–23 1999, pp. 97–101.
 Raymond C.A., Schimleck L.R., Muneri A., Michell A.J.,
Nondestructive sampling of Eucalyptus globulus and E. nitens for
wood properties. III. Predicted pulp yield using near infrared re-
flectance analysis, Wood Sci. Technol. 35 (2001) 203–215.
 Saka S., The relationship between the microfibril orientation in
the tracheid S2 layer and lignin content of coniferous woods, in:
Sudo S. (Ed.) Proceedings of the Pacific Regional Wood Anatomy
Conference, Tsukuba, Ibaraki, Japan, October 1–7, 1984, pp. 100–
 Smook G., Handbook for pulp and paper technologists, Angus
Wilde Publications, Vancouver, BC, 1992.
 Southam C.M., Ehrlich J., Decay resistance and physical character-
istics of wood, J. For. 41 (1943) 666–673.
 TAPPI, Basic density and moisture content of pulpwood, TAPPI
No. T258 om-98, 1989.
 Walker J.C.F., Butterfield B.G., The importance of microfibril angle
for the processing industries, N.Z. For. 40 (1995) 34–40.
 Wallis A.F.A., Wearne R.H., Wright P.J., Analytical characteristics
of plantation eucalypt woods relating to kraft pulp yields, Appita J.
49 (1996) 427–432.
 Wardlaw T., Mohammed C., Barry K., Eyles A., Wiseman D.,
Beadle C., Battaglia M., Pinkard L., Kube P., Interdisciplinary ap-
proach to the study and management of stem defect in eucalypts,
N.Z. J. For. Sci. 33 (2003) 385–398.
 Washusen R., Ades P., Evans R., IlicJ., Vinden P., Relationships be-
tween density, shrinkage, extractives content and microfibril angle
in tension wood from three provenances of 10-year-old Eucalyptus
globulus Labill., Holzforschung 55 (2001) 176–182.
 White D.A., Raymond C.A., Kile G.A., Hall M.F., Are there ge-
netic differences in susceptibility of Eucalyptus nitens and E. reg-
nans stems to defect and decay? Aust. For. 62 (1999) 368–374.
 Wong A.H.H., Wilkes J., Heather W.A., Influence of wood density
and extractives content on the decay resistance of the heartwood
of Eucalyptus delegatensis R.T. Baker, J. Inst. Wood Sci. 9 (1983)
To access this journal online: