Available via license: CC BY 4.0
Content may be subject to copyright.
TROPICS Vol. 33 (2) 119128 Issued December 1, 2024
ORIGINAL ARTICLE
Among-family variations in growth traits, the stress-wave velocity of stems, and
log characteristics of the first-generation Neolamarckia macrophylla (Roxb.)
Bosser in Indonesia
Umi Latifah Dyah Dharmawati1, Ikumi Nezu1, Futoshi Ishiguri1*, Fanny Hidayati2, Agus Ngadianto3,
Yus Andhini Bhekti Pertiwi4, Arif Nirsatmanto5, Sri Sunarti5, Denny Irawati2, Yusuke Takahashi6,
Hikari Yokoyama1, Jyunichi Ohshima1 and Shinso Yokota1
1 School of Agriculture, Utsunomiya University, Utsunomiya 3218505, Japan
2 Faculty of Forestry, Universitas Gadjah Mada, Yogyakarta 55281, Indonesia
3 Vocational College, Universitas Gadjah Mada, Yogyakarta 55281, Indonesia
4 Faculty of Agriculture, Universitas Sebelas Maret, Surakarta 57126, Indonesia
5 National Research and Innovation Agency, Yogyakarta 55281, Indonesia
6 Forest Tree Breeding Center, Forestry and Forest Products Research Institute, Hitachi 3191301, Japan
* Corresponding author: ishiguri@cc.utsunomiya-u.ac.jp
Received: June 6, 2024 Accepted: October 10, 2024
ABSTRACT The growth traits (stem diameter at 1.3 m above the ground and tree height), stress-wave velocity of the stems,
and log characteristics (taper and dynamic Young’s modulus of logs) were examined for 54 trees from 18 half-sib families planted
in a seedling seed orchard of the first-generation Neolamarckia macrophylla (11-year-old) in Wonogiri, Central Java, Indonesia.
The mean values for stem diameter and tree height were 20.2 cm and 20.0 m, respectively. The stress-wave velocity of the stems
was 3.76 km s−1. Meanwhile, the taper and dynamic Young’s modulus of logs were 0.57 cm m−1 and 8.13 GPa, respectively. The
heritability values of each trait were 0.412, 0.365, 0.101, ＜0.001, and 0.092 for the stem diameter, tree height, stress-wave velocity
of stems, taper of logs, and dynamic Young’s modulus of logs, respectively, suggesting that the improvement of all traits is possible
for the next generation, with the exception of the log taper. The 18 half-sib families could be classified into three groups for
different potential uses based on the principal component analysis and cluster analysis results.
Key words: Fast-growing tree species, tree breeding, heritability, dynamic Young’s modulus
INTRODUCTION
The demand for logs as raw materials in the wood
industry in Indonesia has continued to increase over the
years. Erwinsyah et al. (2017) estimated that the production
of logs in Indonesia will increase by 0.61 to 3.32 % every
year if the production of wood-based materials increases by
1 %. In 2022, Indonesian log production reached 64.65
million m3, and almost half of the production was from
Acacia species, which are known as fast-growing tree
species (Livestock, Fisheries, and Forestry Statistics 2023).
This indicates that fast-growing species play a dominant
role and potential in fulfilling the supply of wood demands
in Indonesia. Cossalter and Pye-Smith (2003) reported that
the mean annual increment at an operational scale and time
to reach maturity for the fast-growing tree species in
temperate and tropical regions ranged from 5 to 40 m3 ha−1
year−1, and from 5 to 20 years, respectively. Therefore, logs
from the fast-growing species can be obtained from the
plantation within 20 years for Indonesia’s wood industry.
Aside from Acacia species, several fast-growing
species are found in Indonesia. For instance, Neolamarckia
macrophylla (syn. Anthocephalus macrophyllus (Roxb.)
Havil), known as jabon merah in Indonesian, that naturally
distributed in eastern Indonesia, mainly in Sulawesi and the
Maluku Islands (Irawan and Purwanto 2014). Halawane et
al. (2011) reported that N. macrophylla, which grows
naturally in Sulawesi has an annual radial growth rate of up
to 7 cm year−1 and a height of up to 3 m year−1, indicating
that it can reach 42 cm in stem diameter and 18 m in tree
Copyright © 2024 The Japan Society of Tropical Ecology. This is an open access article distributed under the
terms of Creative Commons Attribution License (CC BY), which permits unrestricted use, distribution, and
reproduction in any medium, provided the original source is properly credited. DOI:10.3759/tropics.MS24-02
Umi Latifah Dyah Dharmawati, Ikumi Nezu et al.
120 TROPICS Vol. 33 (2)
height at the age of six. Based on the data, they also
estimated that the rotation age of plantations of this species
might be 4 to 6 years, when intensive silvicultural treatment
is applied (Halawane et al. 2011). In addition to its fast-
growing characteristics, N. macrophylla has strong
adaptability to various soil types and resistance to pests
(Chung et al. 2016). Recently, these advantages have
resulted in the establishment of plantations to supply wood
for the industry.
To increase the potential of N. macrophylla, the Centre
for Forest Biotechnology and Tree Improvement in
Indonesia has conducted a tree breeding program for this
species. In the program, 55 superior mother trees were
selected based on their growth traits (stem diameter, tree
height, and disease resistance) from Konawe, Southeast
Sulawesi, Indonesia. Using the seeds from the mother trees
selected, a seedling seed orchard was established as the
first-generation trial test site in Wonogiri, Central Java,
Indonesia, in 2012 (Setyaji et al. 2012). After establishing
the test site, several studies concerning growth traits were
conducted at a young age (Surip et al. 2017; Hidayati et al.
2020; Ihda et al. 2023). For example, Surip et al. (2017)
reported that the score for stem straightness in 40-month-
old trees was 2.74, demonstrating good straightness. The
bark thickness of 5-year-old N. macrophylla was 0.36 cm
(Ihda et al. 2023); meanwhile, the stem diameter and tree
height were 18.4 cm and 15.4 m, respectively (Hidayati et
al. 2020). Unfortunately, the wood traits are not yet fully
clarified for the breeding materials of this species. Thus,
wood traits of this species should be clarified to create the
next generation with favorable wood traits for the wood
industry as well as the pulp and paper industry in Indonesia.
In the present study, growth traits (stem diameter at
1.3 m above the ground and tree height) were evaluated in
54 trees from 18 half-sib families planted in a seedling seed
orchard of the first-generation N. macrophylla in Wonogiri,
Central Java, Indonesia. Stress-wave velocity, which is
closely related to Young’s modulus of wood, was also
measured on the stems. In addition, longitudinal variations
of log characteristics (log taper and dynamic Young’s
modulus) were evaluated for the harvested trees. Based on
the results, among-family variations and the heritability of
each trait were esti mated to investigate the possibility of
selecting candidates for the next generation. Furthermore, a
principal component analysis and a cluster analysis were
conducted to investigate the broader utilization of the wood
from this species.
MATERIALS AND METHODS
Field data collection
This study was conducted using the 11-year-old trees
planted in a seedling seed orchard of Neolamarckia
macrophylla (Roxb.) Bosser (syn. Anthocephalus
macrophyllus) located in Wonogiri, Central Java, Indonesia
(07°48′S, 110°54′E; 141 m above sea level; Fig. 1). The
site was the first-generation seedling seed orchard of this
species, with 55 half-sib families originating from Konawe,
Southeast Sulawesi, Indonesia. The 7-month-old seedling
was planted in the seedling seed orchard in February 2012
with randomized complete block designs, including eight
replications (blocks I to VIII). Each family was represented
by four trees in a one-row plot per block. The initial
planting spacing was 4 by 2 m. At age 5, thinning treatment
was conducted in blocks I, II, III, and IV. As a result, one
tree in each family in a block remained. On the other hand,
thinning treatment did not apply to blocks V to VIII. The
climate conditions of the progeny trial site are presented in
Fig. 1.
In July 2023, stem diameter was measured in total of
220 trees (1 tree by 55 families by 4 blocks) in blocks I to
IV. According to the mean (µ) and standard deviation (SD)
of the stem diameter (d) in blocks I to IV, families were
classified into three growth categories: slow growth (d ＜µ
−SD), medium growth (µ−SD≤d＜µ＋SD), and fast
growth (d≥µ＋SD). Six families from each growth
category were selected. Sample trees were then chosen
from the trees of the selected families growing in unthinned
blocks (V to VIII). Among 16 trees (four trees in a one-row
plot per block by four blocks) of each family in four blocks
(V to VIII), three trees were selected based on their stem
diameter ranking: Trees with second to fourth stem diameter
ranking were selected, since the best tree must remain as a
seed source in the seedling seed orchard. As a result, 54
sample trees from 18 families were selected in blocks V to
VIII for measuring the growth and wood traits.
The stem diameter at 1.3 m above the ground and the
tree height were measured for each tree using a diameter
tape (Mejiro, Japan) and a Vertex III instrument (Haglöf,
Sweden), respectively. The stress-wave velocity of the stem
was measured on standing trees using a commercial
handheld stress-wave timer (Fakkop Microsecond Timer,
Fakkop Enterprise, Hungary). According to the methods of
the previous research by Ishiguri et al. (2008), the start and
stop sensors were installed at heights of 1.5 m and 0.5 m
above ground, respectively. The stress-wave propagation
time was measured six times at the same stem position by
Among-family variations of Neolamarckia macrophylla 121
hitting the start sensor with a hammer. The stress-wave
velocity was then calculated by dividing the distance
between two sensors by the average time of the six time
measurements.
The trees were harvested to collect the wood samples
after measuring the tree characteristics. The 2-meter-long
logs were collected from 1.3 m to 7.3 m above the ground
(three logs were obtained from each tree) from harvested
trees. The dynamic Young’s modulus of logs was
determined by the longitudinal vibration method according
to previous researches (Sobue 1986; Ishiguri et al. 2021). A
small hammer hit the cross-section of the butt end to create
sound waves. The first resonance frequency was measured
at the cross-section of the top end using a clip-on
microphone (Fantech, Indonesia) and an application (sound
monitor FFT Wave application, E.N. Software, Japan)
installed in a smartphone (Galaxy A20, Samsung, South
Korea). Before measuring the frequency, the diameters of
both cross ends (including bark) and the length and weight
of the logs were measured using a diameter tape (Mejiro,
Japan), a measuring tape (Joyko, Indonesia), and an
electronic balance (Wesco digital scale, Indonesia),
respectively, to calculate the density of the logs during
testing. The dynamic Young’s modulus (Efr, GPa) of logs
was calculated using the following equation:
Efr＝(2 l f )2 ･ρ･10−9 (1)
where l is the length of the specimen (m), f is the first
resonance frequency (Hz), and ρ is the density of logs at
testing (kg m−3).
Statistical analysis
The statistical analysis was performed using the
statistical software R 4.3.2 version (R Core Team 2023).
Following intercept-only and linear mixed-effects models
with random effects of the family were developed to
evaluate longitudinal variations of log characteristics using
the package lme4 (Bates 2022):
Yj＝µ＋ej (2)
Yjk＝µ＋Familyk＋ejk (3)
Yij＝α0Xij＋α1＋eij (4)
Yijk＝(α0＋Family0k)Xijk＋α1＋eijk (5)
Yijk＝α0Xijk＋α1＋Family1k＋eijk (6)
Yijk＝(α0＋Family0k)Xijk＋Family1k＋α1＋eijk (7)
where Yj is the measured property of the jth individual tree,
Yjk is the measured property of the jth individual tree within
the kth family, Yij is the measured property of the ith height
Fig. 1. Location (a), a photograph (b), and climate data (c) of the
sampling site.
Note: Konawe, Southeast Sulawesi, Indonesia is the loca-
tion of the origin of the mother trees. Wonogiri, Central
Java, Indonesia is the location of the seedling seed orchard
used in the present study. Jakarta is the capital city of
Indonesia (Fig. 1a). The mean monthly temperature and
monthly precipita tion (Fig. 1c) were calculated by
averaging monthly values from 2018 to 2022 (Statistics of
Wonogiri Regency 2023). In Fig. 1c, bars indicate the
mean value of precipitation, and circles indicate the mean
value of the monthly mean temperature.
Umi Latifah Dyah Dharmawati, Ikumi Nezu et al.
122 TROPICS Vol. 33 (2)
position at the top end of the log (m above the ground) of
the jth individual tree, Yijk is the measured property of the
ith height position at the top end of the log (m above the
ground) of the jth individual tree within the kth family, µ is
the grand mean, Xij is the ith height position at the top end
of the log (m above the ground) of the jth individual tree,
Xijk is the ith height position at the top end of the log (m
above the ground) of the jth individual tree within the kth
family, α0 is the fixed slope, α1 is the fixed intercept,
Familyk is the random effect of the kth family, Family0k and
Family1k are the random slope and intercept of the kth
family, and ej,ejk,eij and eijk are residuals. The model with the
minimum Akaike Information Criterion (AIC) (Akaike
1998) was regarded as the optimal model among the
developed models. When the model including the random
effects of the family (Eqs. [3], [5], [6], and [7]) was selected
as the optimal model, the variance component ratio of the
family was calculated using the following formula:
V2
f＝σ2
f / (σ2
f＋σ2
e)･100 (8)
where V2
f (%) is the variance component ratio of the family,
σ2
f is the variance component of the family, and σ2
e is the
residual variance.
To estimate the broad-sense heritability of each trait,
the variance components of the family for the growth and
wood traits were estimated using the following linear
mixed-effects model (9):
yjk＝µ＋Familyk＋ejk (9)
where yjk is the individual mean of the growth and wood
traits of the jth individual tree at the kth family, µ is the
ground mean, Familyk is the random effect of the kth family,
and ejk is residual. The broad-sense heritability of each trait
was estimated by the genetic variance using the following
formula provided by Falconer and Mackay (1989):
H2＝σ2
f / (σ2
f＋σ2
e) (10)
where H2 is broad-sense heritability, σ2
f is the variance
component of the family, and σ2
e is the residual variance.
Using the random effect values of the families
estimated by the model (9), a principal component analysis
(PCA) was performed. A cluster analysis was conducted
using the PCA scores for principal component (PC)1 and
PC2 to categorize the families according to their potential
utilization. Subsequently, the phenotypic correlations (rp)
were determined for the measured traits. In addition, the
genetic correlations (rg) were calculated using random-
effect estimates for family obtained from Eq. (9) by
Pearson’s correlation coefficients.
RESULTS
The mean values of the growth and wood traits of 18
half-sib families Neolamarckia macrophylla are presented
in Tables 1 and 2. The stem diameter ranged from 14.5 cm
in family number 34 to 28.3 cm in family number 70. The
total mean value of the stem diameter from 54 trees was
20.2 cm. The minimum mean value of tree height was
recorded in family numbers 48 and 50 (17.8 m), while the
maximum mean value was in family number 70 (22.9 m).
The total mean value of tree height was 20.0 m.
The stress-wave velocity ranged from 3.60 km s−1 in
family numbers 50 and 59 to 3.94 km s−1 in family number
22, with a total mean value of 3.76 km s−1. The mean values
of the dynamic Young’s modulus of logs ranged from
7.53 GPa in family number 59 to 9.20 GPa in family
number 30, with a total mean value of 8.13 GPa.
Table 2 shows the values of the taper and dynamic
Young’s modulus of logs, which were used to evaluate their
longitudinal variations using mixed-effects models. Based
on the AIC values (Table 3), the intercept-only model (Eq.
[2]) was selected for the longitudinal variation of the log
taper, and the longitudinal variation of dynamic Young’s
Table 1. Mean values of stem diameter, tree height, and stress-
wave velocity of stems in 18 half-sib families.
Family ID nStem diameter
(cm)
Tree height
(m)
Stress-wave
velocity (km s−1)
22 3 21.2 21.2 3.94
26 3 17.0 17.9 3.68
30 3 25.3 21.5 3.85
32 3 18.2 21.4 3.89
34 3 14.5 18.6 3.63
36 3 19.0 18.5 3.87
43 3 20.0 21.6 3.78
48 3 15.0 17.8 3.73
50 3 17.9 17.8 3.60
57 3 19.2 19.0 3.71
58 3 17.8 19.1 3.92
59 3 24.6 21.7 3.60
61 3 20.0 20.8 3.65
65 3 16.4 18.0 3.81
70 3 28.3 22.9 3.66
82 3 22.4 20.9 3.81
103 3 26.9 22.4 3.82
113 3 19.4 18.8 3.77
Minimum 12.9 16.1 3.45
Maximum 54 37.6 25.0 4.17
Mean±SD 20.2±5.0 20.0±2.3 3.76±0.17
Note: n, number of trees; SD, standard deviation.
Among-family variations of Neolamarckia macrophylla 123
modulus was well-fitted to the linear mixed-effects model,
which includes random intercepts of the family (Eq. [6]).
The taper was stable, while the dynamic Young’s modulus
was increased from base to top (Fig. 2). The regression
lines in Fig. 2 were based on the fixed-effect parameters in
the selected models (Table 4).
The variance component and broad-sense heritability
of the growth and wood traits obtained in this study are
presented in Table 5. The heritability values of the stem
diameter, tree height, stress-wave velocity of the stems, and
dynamic Young’s modulus of logs were 0.412, 0.365, 0.101,
and 0.092, respectively. Meanwhile, the heritability value
of the taper was nearly zero.
The phenotypic correlations (rp) between the measured
traits are presented in Table 6. Positive significant correlations
were found between the stem diameter and tree height (rp＝
0.528, p＜0.001) and between the stress-wave velocity of
the stems and the dynamic Young’s modulus of logs (rp ＝
0.352, p＝0.009). However, no correlations were found
between the growth and wood traits.
Similar results were found in terms of genetic correla-
tions (rg) (Table 6). A positive correlation was found between
stem diameter and tree height (rg＝0.844, p＜0.001). The
stress-wave velocity of the stems was also correlated with
Table 2. Mean values of taper and dynamic Young’s modulus of logs in 18 half-sib families.
Family ID n
First log Second log Third log Mean
Taper
(cm/m)
DMOE
(GPa)
Taper
(cm/m)
DMOE
(GPa)
Taper
(cm/m)
DMOE
(GPa)
Taper
(cm/m)
DMOE
(GPa)
22 3 0.26 8.53 1.04 8.54 0.51 8.76 0.61 8.61
26 3 0.79 7.33 0.43 7.65 0.34 8.36 0.52 7.78
30 3 1.08 9.38 0.00 9.14 0.86 9.09 0.65 9.20
32 3 0.70 8.58 0.53 8.89 0.61 9.85 0.61 9.11
34 3 0.35 7.85 0.35 8.45 0.61 8.60 0.43 8.30
36 3 0.17 7.65 0.43 7.99 0.26 8.99 0.29 8.21
43 3 0.56 7.50 0.26 7.57 0.63 8.11 0.48 7.73
48 3 0.78 7.28 0.60 8.17 0.27 7.69 0.55 7.71
50 3 0.95 7.54 0.52 7.81 0.69 8.10 0.72 7.82
57 3 0.69 7.46 0.70 8.18 0.52 7.87 0.63 7.84
58 3 0.60 7.94 0.71 8.30 0.35 7.97 0.55 8.07
59 3 0.54 7.56 0.87 7.39 0.61 7.65 0.67 7.53
61 3 0.96 7.82 0.43 8.62 0.97 9.28 0.79 8.57
65 3 0.95 7.69 0.61 8.17 0.60 7.99 0.72 7.95
70 3 0.54 7.83 0.17 7.55 0.52 7.87 0.41 7.75
82 3 0.35 7.67 0.56 7.71 0.89 8.27 0.60 7.88
103 3 0.82 7.43 0.89 8.40 0.35 8.10 0.69 7.97
113 3 0.17 8.05 0.68 8.23 0.43 8.61 0.43 8.30
Minimum 0.00 6.43 0.00 5.96 0.00 6.63 0.29 7.53
Maximum 54 1.67 11.17 1.91 10.52 2.14 11.21 0.79 9.20
Mean±SD 0.63±0.28 7.84±0.53 0.54±0.26 8.15±0.48 0.56±0.21 8.40±0.60 0.57±0.13 8.13±0.46
Note: n, number of trees; DMOE, dynamic Young’s modulus of logs; SD, standard deviation.
Table 3. Values of Akaike Information Criterion (AIC) in the models for longitudinal variations
of taper and dynamic Young’s modulus of logs.
Trait
Equation
(2) (3) (4) (5) (6) (7)
Taper of logs 165.29 −166.32 −−−
Dynamic Young’s modulus of logs 427.97 421.14 418.84 415.85 414.54 −
Note: Data for taper and dynamic Young’s modulus of logs are at different height positions. The
bold values indicate the selected model with a minimum AIC value; −, the model failed to
converge.
Umi Latifah Dyah Dharmawati, Ikumi Nezu et al.
124 TROPICS Vol. 33 (2)
the dynamic Young’s modulus of logs (rg＝0.509, p＝0.031),
yet no correlations were found between growth and wood
traits.
The results of the principal component analysis of the
random effect of family for the measured traits are presented in
Fig. 3 and Table 7. The cumulative proportion of variance in
PC1 and PC2 was 48.6 % and 35.5 %, respectively (Fig. 3).
High loading values in PC1 were shown for stem diameter
and tree height (0.606 and 0.662, respec tively), whereas
in PC2, the high absolute values were shown for the
stress-wave velocity and dynamic Young’s modulus of logs
(−0.636 and −0.622, respectively) (Table 7). Based on the
Table 4. Parameter estimates, standard errors, t-values, and p-values of fixed-effect parameters and variance
components for a selected model of taper and dynamic Young’s modulus of logs.
Trait Eq.
Fixed-effect parameter Variance component
V2
f (%)
Parameter Estimate SE t-value p-value σ
2
fσ
2
e
Taper 2 µ 0.573 0.032 18.160 ＜0.001 − − −
DMOE 6 α00.142 0.039 3.719 ＜0.001 0.1565 0.6205 20.2
α17.377 0.231 31.890 ＜0.001
Note: DMOE, dynamic Young’s modulus of logs; µ, grand mean; α
0, fixed slope; α
1, fixed intercept; SE,
standard errors; σ
2
f, variance component of family; σ
2
e, residual variance; V2
f (%), variance component
ratio of family.
Fig. 2. Longitudinal variations of taper (a) and dynamic Young’s
modulus (b) of logs in 54 individual trees of 18 half-sib
families.
Note: Regression lines were based on the fixed-effects
parameters in the selected model in Table 4.
Table 5. Variance components and broad-sense heritability in each measured trait.
Trait
Variance component
Heritability
σ
2
fσ
2
eTotal
Stem diameter 10.530 15.050 25.580 0.412
Tree height 1.940 3.372 5.312 0.365
Stress-wave velocity of stems 0.003 0.027 0.030 0.101
Taper of logs ＜0.001 0.057 0.057 ＜0.001
Dynamic Young’s modulus of logs 0.053 0.524 0.577 0.092
Note: ; σ
2
f, variance component of family; σ
2
e, residual variance.
Table 6. Correlations between measured traits.
Trait 1 Trait 2
Phenotypic Genetic
rpp-value rgp-value
D TH 0.528 <0.001 0.844 <0.001
SWV −0.129 0.353 0.029 0.901
DMOE −0.151 0.275 0.021 0.933
TH SWV −0.106 0.444 0.151 0.550
DMOE 0.126 0.364 0.217 0.387
SWV DMOE 0.352 0.009 0.509 0.031
Note: D, stem diameter; TH, tree height; SWV, stress-wave velocity
of stems; DMOE, dynamic Young’s modulus of logs; rp, pheno
typic
correlation coefficient; rg, genetic correlation coefficient. Number of
trees in phenotypic correlations is 54. Number of families in genetic
correlations in the Eq. (9) is 18.
Among-family variations of Neolamarckia macrophylla 125
principal component analysis scores, 18 half-sib families
were classified into three groups (Fig. 3), and the potential
utilization of each group was examined in Table 8.
DISCUSSION
Growth and wood traits
The annual radial growth rates (stem diameter / tree age)
in the present study ranged from 1.32 cm year−1 (14.5 cm /
11 years) to 2.57 cm year−1 (28.3 cm / 11 years), lower than
Fig. 3. Principal component analysis plot (a) and dendrogram (b) based on the principal component scores.
Note: PC, principal component. Numbers in Figs. indicate family IDs. Alphabets indicate cluster groups. Boxes indicate families
in the same group.
Table 7. Loading values of principal component analysis for
18 half-sib families.
Trait PC1 PC2
Stem diameter 0.606 0.392
Tree height 0.662 0.235
Stress-wave velocity of stems 0.299 −0.636
Dynamic Young’s modulus of logs 0.326 −0.622
Note: PC, principal component.
Table 8. Characteristics of each group.
Cluster Growth trait Wood trait Description Expected utilization
A＋＋＋ Medium wood traits but poor growth traits Solid wood production for non-structural use
B＋＋＋ ＋＋＋ Superior families both growth and wood
traits
Candidate of the next generation producing
solid wood for structural use
C＋＋＋ ＋＋ Good growth traits with lower wood traits Biomass production, veneer production, and
pulp and paper production
Note: ＋＋＋, very good;＋＋, average;＋, poor; cluster A to C refer to Fig. 3.
Umi Latifah Dyah Dharmawati, Ikumi Nezu et al.
126 TROPICS Vol. 33 (2)
the annual radial growth rate of N. macrophylla from the
same seedling seed orchard at 3 years old, which was
4.4 cm year−1 (13.2 cm / 3 years) (Surip et al. 2017); and
also 5-year-old trees reported by Hidayati et al. (2020),
2.70 cm year−1 (13.5 cm / 5 years) to 4.30 cm year−1
(21.5 cm / 5 years). In another study, the annual radial
growth rate of 4-year-old trees grown in Gorontalo was
3.32 cm year−1 (13.3 cm / 4 years) to 3.64 cm year−1
(14.6 cm / 4 years) (Sandalayuk et al. 2023). These results
suggest that the annual radial growth rate in the present
study is slower than that in the previous studies. On the
other hand, the radial growth rate is usually not equal
during tree life: The rate appeared fast at younger ages, and
then became slow as tree age increased (Salas-Eljatib et al.
2021; Nezu et al. 2023). In 12-year-old Eucalyptus
camaldulensis planted in Thailand, the mean annual
increment gradually decreased from younger ages up to 12
years of age (Nezu et al. 2023). Considering that the
expected harvesting period in N. macrophylla is 4 to 6
years, the radial growth rate after 4 to 6 years of planting
might become slow, leading to the slower annual radial
growth rate obtained in the present study (1.32 cm year−1 to
2.57 cm year−1 for 11-year-old trees) compared to other
previous research results. Further research is needed to
clarify the age-related changes in the radial growth rate of
this species.
Compared to other fast-growing species, the dynamic
Young’s modulus of logs in the present study was lower
than for Acacia mangium (11.70 GPa, Ngadianto et al.
2020) and Eucalyptus camaldulensis (11.72 GPa, Ishiguri et
al. 2013), although the stress-wave velocity of stems was
similar (A. mangium＝3.60 to 3.77 km s−1, Masendra et al.
2023; E. camaldulensis＝3.45 km s−1, Ishiguri et al. 2013).
These differences might be due to differences in measuring
methods between the dynamic Young’s modulus of logs
and the stress-wave velocity of stems, and in radial
variations of wood properties, such as wood density and
microfibril angle (MFA). On the other hand, Pertiwi et al.
(2017) reported that the stress-wave velocity of stems and
the dynamic Young’s modulus of logs in 4-year-old
Neolamarckia cadamba trees planted in Indonesia were
2.99 km s−1 and 6.54 to 6.64 GPa, respectively. Thus,
Young’s modulus of logs and the stress-wave velocity of
stems in N. macrophylla might have a higher value than
those in N. cadamba (Pertiwi et al. 2017), although the tree
age differed.
The broad-sense heritability of stem diameter and tree
height in the present study (Table 5) was similar to the
heritability values at the family level for 55 half-sib families
of N. macrophylla (0.463 for stem diameter and 0.213 for
tree height; Surip et al. 2017). The heritability of growth
traits in this study showed similar or relatively higher
values compared to the previous study (Surip et al. 2017).
Thus, growth characteristics can be improved through the
selection of superior families under appropriate tree
breeding programs. Although the heritability showed lower
values for stress-wave velocity and dynamic Young’s
modulus of logs than growth characteristics, the heritability
was around 0.1, suggesting that wood traits also might be
genetically improved through family or within-family
selection in this species.
Longitudinal variations of log characteristics
The selected model explaining the longitudinal
variation of the log taper was an intercept-only model (Eq.
[2], Table 3), indicating that the taper values were almost
stable from the base to the top (Fig. 2). In general, trees
show that forcing the crown up the stem by pruning
decreases the taper (Punches 2004). Halawane et al. (2011)
reported that N. macrophylla has self-pruning abilities.
Thus, this ability might reduce the taper due to the uniform
growth along the stem of this species. On the other hand,
the selected model in the taper (Eq. [2]) did not include the
random effect of family. These results suggest that the stem
form of the N. macrophylla tree is cylindrical, despite the
differences between the families. In addition, the heritability
of the log taper was almost zero (Table 5). Based on these
results, the log taper of this species might be improved by
the silvicultural treatments, such as by controlling the
crown size by pruning and changing the tree spacings.
The fixed slope α0 was 0.142 in the selected model of
the dynamic Young’s modulus of logs (Table 4), suggesting
that the dynamic Young’s modulus slightly increased from
the base to the top in N. macrophylla (Fig. 2). Young’s
modulus of wood is closely related to the wood density
(Saranpää 2003) and the MFA in the S2 layer of tracheids
and wood fibers (Donaldson 2008). Thus, the increase of
the dynamic Young’s modulus of logs in this study might
be due to a higher wood density or a lower microfibril angle
(MFA) along the stem. Further research is needed to clarify
the effects of wood density and MFA on Young’s modulus
of wood and on radial variations of wood density and MFA
in N. macrophylla.
The selected model explaining the dynamic Young’s
modulus of logs (Eq. [6], Table 3) included a random
intercept of the family, suggesting that the ranking at the
family levels in the dynamic Young’s modulus of logs
might be the same at each height position. In addition, the
Among-family variations of Neolamarckia macrophylla 127
random slope of the family was not included in Eq. (6) for
the dynamic Young’s modulus, indicating that the
increasing ratio from the base to the top might be almost
the same among families. Furthermore, the variance
component of the family was moderate (ca. 20 % in Table
4). Therefore, selecting the families with a high dynamic
Young’s modulus in the first logs might lead to the effective
utilization of the N. macrophylla wood as solid wood.
Relationships between traits
Table 6 indicates a positive correlation between the
stress-wave velocity of stems and the dynamic Young’s
modulus of logs. The same results were reported in previous
studies (Ishiguri et al. 2008, 2021): The stress-wave
velocity of stems has a positive correlation with the log’s
dynamic Young’s modulus. Thus, the stress-wave velocity
on a standing tree can predict the Young’s modulus of a log,
even if the stress-wave velocity might only be measured at
outer parts of the stems.
The positive correlations, both phenotypic (rp) and
genetic (rg), between the stem diameter and tree height and
between stress-wave velocity of stems and the dynamic
Young’s modulus of log might provide a more straight-
forward selection process. Considering their heritability, the
stem diameter and stress-wave velocity of stems are easier
to measure in the field. Thus, these two traits (stem diameter
and stress-wave velocity of stems) can be used as selection
criteria to increase genetic gain.
Characterization of families
According to Halawane et al. (2011), N. macropylla is
commonly used as plywood, pulp and paper, boards, crates,
furniture, and non-construction building materials. In the
present study, based on the cluster analysis according to the
principal component analysis (Table 7 and Fig. 3), the
families were divided into three groups. The characteristics
and potential utilization of each group are presented in
Table 8. The families in group A exhibited average values
in wood traits but had below-average growth traits. The
wood from families in group B is considered superior
because they have the best growth and wood traits among
the other families. Thus, the families within group B may
be candidates for the next generation to improve the growth
and wood traits of N. macrophylla. The wood within these
families can be used for solid wood production for
structural purposes. On the other hand, group C has very
good growth traits. The wood within the families in group
C is suitable as raw material for pulp and paper or biomass
products. In addition, the moderate decrease in diameter of
this species suggests that the stem shape is relatively
cylindrical based on its taper values (Table 2), allowing this
wood to be utilized as a raw material for veneer because it
can produce a high yield.
CONCLUSION
The growth traits, stress-wave velocity of the stem,
and dynamic Young’s modulus of logs were evaluated for
18 half-sib families of the 11-year-old first-generation N.
macrophylla grown in Central Java, Indonesia. Positive
significant relationships were found between stem diameter
and tree height, and between the stress-wave velocity of
stems and the dynamic Young’s modulus of logs. Thus, the
stem diameter and stress-wave velocity on standing trees
can be powerful tools for selecting the criteria of superior
trees in the tree breeding program for this species. No
significant correlation was found between growth traits and
wood traits. According to broad-sense heritability and the
correlation between measured traits, both growth and wood
traits can be improved simultaneously by selecting superior
families based on these traits. On the other hand, the shape
of the logs might be improved by effective silvicultural
treatments. Based on the PCA and cluster analysis, 18
families were classified into three groups: Group A, non-
structural solid wood; Group B, candidates for next-
generation and solid wood for structural purposes; and
Group C, for production purposes requiring high wood
yields. The obtained results in the present study can
contribute to the increase in N. macrophylla wood volume
with desirable properties for the wood industry.
ACKNOWLEDGMENTS
The authors would like to express their appreciation to
The Centre for Forest Biotechnology and Tree Improvement,
Indonesia, and the Environmental Agency of Wonogiri
Regency, Indonesia, for providing the wood samples. They
would also like to extend their appreciation to the students
of the Department of Forest Products Technology, Faculty
of Forestry, Universitas Gadjah Mada, for their help in the
field experiments.
Umi Latifah Dyah Dharmawati, Ikumi Nezu et al.
128 TROPICS Vol. 33 (2)
REFERENCES
Akaike H. 1998. Information theory and an extension of the
maximum likelihood principle. In: Parzen E, Tanabe K,
Kitagawa G, editors. Selected papers of Hirotugu Akaike.
Springer series in statistics. New York: Springer; p. 199213.
Bates DM. 2022. lme4: Mixed-effects modelling with R. https://
people.math.ethz.ch/~ maechler/MEMo-pages/1MMwR.pdf
(cited February 2, 2024).
Chung AYC, Pereira JT, Hastie A, Paul V, Johnlee EB. 2016. New
records of insects associated with red laran (Neolamarckia
macrophylla) in Sabah. Sepilok Bulletin 23&24: 114.
Cossalter C and Pye-Smith C. 2003. Fast-wood forestry: Myths and
realities. Center for International Forestry Research. Bogor.
Donaldson L. 2008. Microfibril angle: measurement, variation, and
relationships – A review. IAWA Journal 29: 345386.
Erwinsyah, Harianto, Sinaga BM, Simangunsong BCH. 2017.
Analisis penawaran dan permintaan kayu bulat untuk pemenuhan
kebutuhan industri kayu lapis, kayu gergaji dan pulp di
Indonesia [Analysis of supply and demand of logs for
plywood, sawn timber and pulp industries in Indonesia].
Sosio-E-Kons 9: 117124.
Falconer DS, Mackay TFC. 1989. An introduction to quantitative
genetics 3rd edition. Longman Scientific and Technical. Essex.
Halawane JE, Hidayah HN, Kinho J. 2011. Prospek pengembangan
jabon merah (Anthocephalus macrophyllus (Roxb.) Havil),
solusi kebutuhan kayu masa depan [Jabon merah (Anthocephalus
macrophyllus (Roxb.) Havil) development prospects, solutions
to future challenges]. Research and Development Institute for
Environment and Forestry Manado. North Sulawesi.
Hidayati F, Sunarti S, Setyaji T, Nirsatmanto A. 2020. Physical and
mechanical properties of jabon merah (Anthocephalus
macrophyllus) planted in Wonogiri, Central Java. Journal of
Tropical Forests 8: 357365.
Ihda FV, Lukmandaru G, Nirstamanto A. 2023. Extractive content
of white jabon (Neolamarckia cadamba) and red jabon
(Neolamarckia marophyllus) stemwood from Wonogiri,
Central Java. Jurnal Ilmu Kehutanan 17: 6775.
Irawan US, Purwanto E. 2014. White jabon (Anthocephalus
cadamba) and red jabon (Anthocephalus macrophyllus) for
community land rehabilitation: improving local propagation
efforts. Agricultural Science 2: 3645.
Ishiguri F, Aiso H, Aruga K, Ohshima J, Iizuka K, Yokota S. 2021.
Stress-wave velocity of stems, dynamic Young’s modulus of
logs, and wood properties in two broad-leaved tree species.
Bulletin of the Utsunomiya University Forests 57: 1521.
Ishiguri F, Diloksumpun S, Tanabe J, Iizuka K, Yokota S. 2013.
Stress-wave velocity of trees and dynamic Young’s modulus
of logs of 4-year-old Eucalyptus camaldulensis trees selected
for pulpwood production in Thailand. Journal of Wood
Science 59: 506511.
Ishiguri F, Matsui R, Iizuka K, Yokota S, Yoshizawa N. 2008.
Prediction of the mechanical properties of lumber by stress-
wave velocity and Pylodin penetration of 36-year-old
Japanese larch trees. Holz als Roh-und Werkstoff 66: 275280.
Livestock, Fisheries, and Forestry Statistics. 2023. Statistics of
forestry production 2022. Statistics Indonesia. Jakarta.
Masendra, Nezu I, Ishiguri F, Hidayati F, Nirsatmanto A, Sunarti S,
Surip, Kartikaningtyas D, Takashima Y, Takahashi Y, Ohshima
J, Yokota S. 2023. Variations of growth and wood traits in
standing trees of the third-generation Acacia mangium
families in Indonesia. Silvae Genetica 72: 150162.
Nezu I, Ishiguri F, Aiso H, Diloksumpun S, Ohshima J, Yokota S.
2023. Evaluation of xylem maturation manner for solid wood
production from Eucalyptus camaldulensis planted in
Thailand. Tropics 32: 4963.
Ngadianto A, Ishiguri F, Nezu I, Takahashi Y, Tanabe J, Hidayati F,
Irawati D, Ohshima J, Yokota S. 2020. Wood properties and
simulated modulus of elasticity of glulam in three fast-
growing tree species grown in community forests in
Yogyakarta, Java Island, Indonesia. Tropics 29: 89104.
Pertiwi YAB, Aiso H, Ishiguri F, Wedatama S, Marsoem SN,
Ohshima J, Iizuka K, Yokota S. 2017. Effect of radial growth
rate on wood properties of Neolamarckia cadamba. Journal of
Tropical Forest Science 29: 3036.
Punches J. 2004. Tree growth, forest management, and their
implications for wood quality. Pacific Northwest Extension
Publication (PNW 576). Oregon State University.
R Core Team. 2023. R: A language and environment for statistical
computing. R Foundation for Statistical Computing, Vienna,
Austria. https://www.r-project.org/ (cited November 2, 2023).
Salas-Eljatib C, Mehtätalo L, Gregoire TG, Soto DP, Vargas-Gaete
R. 2021. Growth equations in forest research: Mathematical
basis and model similarities. Current Forestry Report 7: 230244.
Sandalayuk D, Masrukhin, Djabar M, Ruruh A, Ernikawati. 2023.
Thinning on the growth of jabon merah trees (Anthocephalus
macrophyllus) in the industrial plantation forest area. Journal
of Research in Science Education 9: 58525858.
Saranpää P. 2003. Wood density and growth. In: Barnett J and
Jeronimidis G (eds) Wood quality and its biological basis.
Blackwell Publishing, Oxford. 87113.
Setyaji T, Sunarti S, Nirsatmanto A, Kartikaningtyas D, Yuliastuti
DS. 2012. General information of seed source (F–1) of
Anthocephalus cadamba and Anthocephalus macrophyllus
establishment in Wonogiri, Central Java Fiscal Year 20102012.
FTIP-P3-No.60. Forest Research and Development Agency.
Ministry of Forestry Indonesia.
Sobue N. 1986. Measurement of Young’s modulus by the transient
longitudinal vibration of wooden beams using a fast Fourier trans-
formation spectrum analyzer. Mokuzai Gakkaishi 32: 744747.
Statistics of Wonogiri Regency. 2023. https://wonogirikab.bps.
go.id/indicator/151/164/1 /curah-dan-hari-hujan.html (cited
December 7, 2023).
Surip, Indrioko S, Nirsatmanto A, Setyaji T. 2017. Effect of
selection on genetic gain of first-generation seedling seed
orchards of jabon merah (Anthocephalus macrophyllus
(Roxb.) Havil) established in Wonogiri. Jurnal Pemuliaan
Tanaman Hutan 11: 3344.
