全 文 ：Journal of Forestry Research (2009) 20(3): 261−267
DOI 10.1007/s11676-009-0045-8





Geographic variation and provenance selection for bamboo wood prop-
erties in Bambusa chungii

YANG Xiu-yan1, 2, FU Mao-yi2*, XIE Jin-zhong2, LI Zheng-cai2
1 Research Institute of Forestry, Chinese Academy of Forestry, Beijing 100091, P. R. China
2 Research Institute of Subtropical Forestry, Chinese Academy of Forestry, Fuyang 311400, P. R. China

Abstract: Using 3-year-old culms of 8 provenances of Bambusa chungii from Guangdong, Guangxi and Hainan Provinces, the indexes of
wood properties, such as fiber dimensions and chemical composition were investigated and analyzed by the methods of Analysis of variance
and correlation coefficient to reveal the geographic genetic variation situation. The results showed that there are significant differences be-
tween fiber length, fibrin and 1% NaOH extraction contents of B. chungii from 8 provenances; moreover, the fiber length and fiber
length/width ratio had a decreasing change pattern with geographic variation from the south to the north in altitude (from high to low in ele-
vation). The heritabilities for fibrin, fiber length, 1% NaOH extractive, lignin and fiber length/width ratio were 0.7, 0.84, 0.54, 0.38 and 0.13,
respectively. A significant negative correlation was found between 1% NaOH extraction, benzo-alcohol extraction contents and bamboo
culm yield, whereas there was a significant positive correlation between fibrin contents and bamboo wood yield. Besides, a close correlation
was detected between fiber dimensions indices and bamboo growth or bamboo wood yield. Finally, three provenances with high qualities
and culm yield, i.e. Huaiji, Xinyi and Guilin, were selected as superior sources based on principal component analysis.
Keywords: Bambusa chungii; bamboo wood property; pulp-bamboo; geographic variation; provenance selection



Introduction

Paper industry has been developing rapidly in China. At present,
China ranks in the third place at world level for total paper and
paper-board production, only behind the United State of America
and Japan (Gan 2002). However, the paper industry of China is
facing a challenge due to the shortage in raw materials, which
has forced major imports from foreign countries. It is difficult for
China to develop the paper industry only depending on the wood
pulp stands because of the shortage of annual wood stock output
(Ma et al. 2004). Fortunately, China is rich in bamboo re-
sources. The indexes of bamboo wood fiber and chemical com-
ponents show that bamboo wood has a good pulp potential. The
bamboo wood can be used to make bamboo pulp or bam-

Foundation project: This research was funded by both of the Interna-
tional Tropical Timber Organization (ITTO), grant No. PD10/00. REV.
2 (I,F), and the National ‘Tenth Five-year’ Key Program of China,
grant No. 2004BA506B0103.
Received: 2008-12-15; Accepted: 2009-02-10
© Northeast Forestry University and Springer-Verlag 2009
The online version is available at http://www.springerlink.com
Biography: YANG Xiu-yan (1973-), female, Postdoctoral Fellow in
Research Institute of Forestry, Chinese Academy of Forestry, Beijing
100091, P. R. China. (E-mail:sueyxy@126.com)
*Corresponding author: Email: E-mail: fumy1@163.net
Responsible editor: Zhu Hong

boo-wood mixed pulp to produce various types of paper. Using
bamboo as partially substitute for wood could be an effective
way for the paper industry to overcome the shortage of wood
pulp.
Many studies showed that different bamboo species have ob-
vious differences in bamboo wood properties and chemical
components. Most bamboos, especially sympodial bamboos, are
superior materials for papermaking (Ma et al. 2004; Wang et al.
1999; Hui et al. 1993). Bambusa chungii, belonging to the Bam-
busa genera, is an important sympodial bamboo species in the
south of China, being distributed mainly in Guangdong, Guangxi,
Hainan, and Fujian provinces. B. chungii has high economic
values and can be used as raw material for strip and pa-
per-making. Xinyi county of Guangdong Province, with an area
of 20 ×103 ha, has the largest distribution of B. chungii.
Studies on wood properties in wood plants had demonstrated
that there were abundant genetic variations at different genetic
control levels (Wright et al. 1993; Matheson et al. 1986). How-
ever, few studies have been reported on bamboo wood properties,
although bamboo has multi uses and high economic values. In
the present study, the properties of fiber dimensions and chemi-
cal components in bamboo wood were investigated from 8
provenances of B. chungii with 3-year-old culms. The objective
of this study was to find out the geographic genetic variations in
bamboo wood properties and growth traits among the various
provenances, and the relationship between genetic variation and
their ecological environment factors, so as to improve the effec-
tive utilization of the bamboos and to give some useful informa-
tion to bamboo planters.
RESEARCH PAPER
Journal of Forestry Research (2009) 20(3): 261−267

262

Materials and methods

Bamboo wood material

Bamboo wood samples used in this study were taken from a B.
chungii provenance test station established with 8 provenances of
B. chungii at Maoming, Guangdong, in the south of China. In
accordance with the natural distribution of B. chungii, all sam-
ples of the provenances were collected from 8 natural popula-
tions in Guangdong, Hainan and Guangxi provinces. Detailed
information of the sampling plots is given in Table 1. The dis-
tance between two populations was more than 200 km to avoid
sampling mother bamboo from the same population, and the
distance between two samples was 50 m apart. Mother bamboos
were planted at 3 m×4 m spacing in 2001. The experiment was
arranged in three randomized complete blocks and each prove-
nance was established in a 12-row clump plots. There were two
rows of the same species from local populations as fence sur-
rounding the trial site. No particular silvicultural treatment was
performed before the collection of the timber sample.

Table 1. Characteristics of site and climate of eight sampling plots of Bambusa chungii
Provenance Longitude Latitude
Aaltitude
(m)
Annual mean
temp. (ºC)
Minimum
temp.
（ºC）
January monthly
mean temp.
（ºC）
July monthly
mean temp.
（ºC）
Annual
precipitation
（mm）
Relative
humidity
（%）
Frostless
days
（d）
Qiongshan, Hainan 110º30′ 19º50′ 35 23.8 2.8 12.1 28.3 1676.4 85 365
Nanning, Guangxi 108º21′ 22º49′ 90 21.6 -2.1 12.8 28.2 1304.2 79 360
Guilin, Guangxi 110º30′ 24º46′ 135 18.9 -2.1 15.6 23.0 1949.5 79 309
Fengkai, Guangdong 111º33′ 23º30′ 25 20.8 -3.4 11.2 28.6 1420.8 82 355
Huaiji, Guangdong 112º12′ 23º50′ 90 20.8 -3.9 13.5 28.8 1785.4 80.2 310
Xinyi, Guangdong 110º55′ 22º21′ 167 22.8 -1.8 13.0 28.0 1800.0 85 340
Zengcheng, Guangdong 113º52′ 23º17′ 56 21.2 -2.2 13.2 28.2 1967.2 78 359
Qingyuan, Guangdong 113º03′ 23º31′ 100 21.6 -0.6 12.4 28.8 2173.6 78 351

Environmental conditions of study area

The test site was located in Maoming Forest Park (latitude
21º39´ N, longitude 110º50´E, with an elevation of 23 m above
sea level), under south subtropical monsoon climate. The mean
annual temperature, minimum temperature,. mean temperature of
January, and the mean temperature of July are 23ºC, 2.8ºC,
16.1ºC, and 28.5ºC, respectively. The mean annual precipitation
in this area is 1704.9 mm, with relative humidity of 81%. The
annual evaporation ranged from 1808.2 to 1993.7 mm; therefore
the site has regular droughts every year. The frostless period is
over 350 days in the test site. The soil belongs to laterite soil
types.
Sample collection of bamboo wood

Six clumps per provenance and sixteen clumps per block were
chosen for the study. Totally, 48 culms were felled at the ground
level and 10 cm of their top part was cut down. Three bamboo
segments (about 40 cm long) from three parts, i.e. at the base, 1.3
m (the standard breast height) above ground level, and top part,
were collected from each culm. Selected culms and clumps were
free of any physical and mechanical damage and were not in-
fected by any plagues and diseases. All the samples were marked
and DBH, height of culm, culm number per clump, clear height
and node number under the first branch of the sample culms
were measured (Table 2).

Table 2. Information of bamboo samples
Thickness of bamboo wall (mm)
Provenance
Height
(m)
DBH
(cm)
Culms per
clump
Clear height
(m)
Node num-
bera Culm base DBH Culm top
Fengkai, Guangdong 4.58 2.28 24.3 3.7 6.9 0.52 0.34 0.13
Guilin, Guangxi 4.92 2.42 33.9 2.9 6.9 0.52 0.34 0.13
Huaiji, Guangdong 4.87 2.57 32.3 3.1 6.5 0.44 0.28 0.12
Qiongshan, Hainan 3.98 2.13 23.7 2.6 6.7 0.45 0.28 0.11
Nanning, Guangxi 4.95 2.63 31.4 3.1 7.2 0.52 0.35 0.10
Qingyuan, Guangdong 3.70 1.87 56.4 2.5 5.9 0.44 0.26 0.11
Xinyi, Guangdong 4.66 2.62 31.3 3.2 7.4 0.51 0.34 0.12
Zengcheng, Guangdong 4.73 2.52 40.2 2.9 6.7 0.45 0.28 0.12
Notes: “a” is number of nodes below the first branch.

Fiber dimensions of bamboo wood

The sample was boiled several times to break down the fibers. A
mixed liquor of hydrogen peroxide (30%−35% concentration)
and glacial acetic acid with ratio of 1:1 was used to macerate
match-stick pieces from the sample. The unbroken fibers were
measured by Kajanni FS-100 fiber automatic analyzer and pho-
tos were taken by WV-CP470 microscope camera.
Journal of Forestry Research (2009) 20(3): 261−267

263

Chemical components of bamboo wood

The bamboo wood samples were cut into pieces and milled into
powder to pass through 0.45-μm sieve for analysis. 1% NaOH
extraction content, benzo-alcohol extraction content and lignin
were measured following the national standards methods
GB2677.5-81, GB2677.7-81 and GB2677. 8-81. Cellulose was
measured by the Nitric acid-ethanol method.

Yield measurement of bamboo wood

Thirty culms with normal growth and without any physical and
mechanical damage were selected. From ground level, DBH and
weight of each culm were determined. The data were fit to a
Yield equation using a regression model.

Statistical analyses

A mixed linear model was used to estimate variance components.

ijkjiijk eFRY +++= μ (1)

where, Yijk is a phenotypic individual observation; μ is the overall
mean; Ri is the fixed replication effect; Fj is the random prove-
nance effect with E(Fj)=0 and var(Fj)=σf2; and eijk is the random
residual effect with E(eijk)=0 and var(eijk) =σe2.
The broad-sense provenance heritability (H2) was calculated
for each trait measured by the following equation.

2222 /)( fefH σσσ −= (2)

The data were analyzed using SAS 8.0. ANOVA. Pearson SAS
procedures were used to analyze the variance of each trait. and
the correlation coefficients between different traits based on
mean value of the data of three parts of the culm (including base,
DBH and top of the culm) were calculated. The data were nor-
malized using an arcsine transformation before the statistics
analyses.
The power function equation was chosen for fitting the rela-
tionship between culm yields(y, fresh weight (kg)) and DBH (X).
The regression equation of culm fresh weight-DBH was

y=0.1655X2.1481(p<0.001, R=0.9442) (3)

The output of a clump could be calculated multiplying the weight
of a single culm by the number of culms per clump.

Results and analysis

Variation of fiber dimensions

ANOVA for fiber dimensions and chemical compositions were
analyzed in this study (Table 3). Fiber dimensions directly im-
pact the techniques of papermaking and pulp performance. Gen-
erally, bamboo wood with fiber of longer length, better flexibility
and higher ratio of fiber length to width is regards as better raw
material for papermaking. The average length of fiber of B.
chungii from all the provenances was over 4 mm, which was
longer than that of other bamboos (2.09 mm) and broadleaf spe-
cies (1.04 mm), and even longer than that of some conifers (Qin
2003). The ratio of fiber length to width was very high in B.
chungii. The range of B. chungii fiber length from different
provenances varied from 4.25 mm to 4.65 mm, with a significant
(p<0.01) difference due to a small coefficient of variation, about
2.3%. No significant differences in fiber width and ratio of fiber
length to width of B. chungii were found among the provenances.
However, significant differences between the two trait indexes
(fiber width and ratio of fiber length to width) were observed
among all individuals and the coefficients of variance (COV)
were 19.56% and 16.77% in Qiongshan, Hainan population,
respectively. These results suggest that it may be more effective
to perform individual selection on fiber dimensions.

Table 3. ANOVA for chemical compositions and fiber dimension of bamboo wood
Index Mean Range CV(%) Std Dev F value H 2
1%NaOH extractive(%) 25.30 23.7-26.46 4.95 1.25 2.18+ 0．54
Benzo-alcohol extractive (%) 5.99 5.61-6.44 12.87 0.77 0.37 a
HNO3-alcohol fibrin (%) 46.23 43.68-50.87 4.42 2.04 3.46* 0．71
Lignin(%) 22.73 21.31-23.94 5.81 1.32 1.60 0．38
Fiber length(mm) 4.42 4.25-4.65 2.37 0.105 6.34** 0．84
Fiber width(μm) 13.34 12.73-14.08 12.13 1.62 0.27 a
Ratio of L/W 334.92 311.67-373 11.31 37.86 1.15 0．13
Notes: **,*,+ indicate statistically significant at 0.01, 0.05 and 0.1 level respectively. a: indicates H 2 for that index can not be estimated.

Variations of chemical components of bamboo wood

Chemical components of bamboo wood have important effect on
the quality and performance of paper products. In other words, a
higher fibrin, lower lignin and lower extraction would benefit the
yield and quality of pulp. Variance analysis for chemical com-
ponents demonstrated that there are significant differences be-
tween fibrin content and 1% NaOH extraction at 5% and 10%
level, respectively, but there are no significant differences in
lignin and benzo-alcohol extractive among the eight provenances.
Journal of Forestry Research (2009) 20(3): 261−267

264
Further analysis showed that the fibrin content of bamboo wood
fluctuated from 43.68% to 50.87% among the 8 provenances.
This change (greater than 7%) indicates that there is a potential
for provenance selection. Additionally, the absolute variation
value of lignin content was small among provenances and indi-
viduals, and benzo-alcohol extract had a large difference among
individuals (Table 4). The content of 1% NaOH extract, fibrin
and benzo-alcohol extract follows a normal distribution (Fig. 1)
indicating that it is also possible to select superior clump with
good chemistry for further propagation and utilization.

Fig. 1 Frequency distribution of some chemistry properties of Bambusa chungii

Table 4. Variation of bamboo wood properties of B. chungii within provenances
Fiber length
Fiber width
Ratio of L/W 1%NaOH
extractive
Benzo-alcohol
extractive
Fibrin Lignin
Provenance.
Mean
(mm)
CV
(%)
Mean
(μm)
CV
(%)
Mean
(%)
CV
(%)
Mean
(%)
CV
(%)
Mean
(%)
CV
(%)
Mean
(%)
CV
(%)
Mean
(%)
CV
(%)
Fengkai, Guangdong 4.63 1.92 12.94 11.75 361 12.52 26.31 3.65 5.90 16.1 44.72 1.68 22.26 6.29
Guilin, Guangxi 4.31 0.58 13.74 12.88 316 11.85 25.68 5.72 6.44 7.45 47.69 4.28 22.58 5.36
Huaiji, Guangdong 4.48 4.70 13.26 17.27 343 12.34 23.81 7.43 5.65 16.81 50.87 3.09 23.87 1.68
Qiongshan, Hainan 4.65 2.57 12.73 19.56 373 16.77 26.46 1.55 6.01 11.65 45.5 3.74 23.60 4.87
Nanning, Guangxi 4.38 1.47 14.08 10.94 313 9.54 25.26 2.34 6.12 20.59 46.16 1.73 22.21 4.59
Qingyuan, Guangdong 4.25 0.00 13.67 5.71 312 5.74 23.70 3.88 5.61 3.74 43.68 2.95 23.94 6.22
Xinyi, Guangdong 4.35 2.79 12.81 6.56 341 8.99 26.10 6.51 6.07 8.24 45.75 3.98 21.31 7.84
Zengcheng, Guangdong 4.31 0.94 13.46 4.23 320 4.06 25.04 5.79 6.13 9.62 45.48 9.26 22.05 7.85


Trend in geographical variation of bamboo wood properties

The correlation analysis between fiber dimensions, chemical
components, and geographic and climate factors showed that
fiber length became shorter with an increase in altitude (Table 5).
The ratio of length to width (L/W) became lower with the in-
crease of the latitude. There were significant negative correla-
tions (α=0.1) between fiber dimensions, altitude and latitude, but
the other traits did not follow any geographic trend. These results
were in aggreement with previous studies. Ma and Zhu (1990)
found that fiber length and ratio of L/W of B. chungii in the
south of Zhejiang Province were smaller than those in low lati-
tude regions. The variation of fiber dimension with latitude is an
adaptation to its physical support function, i.e. the faster growing
bamboo of lower latitudes (altitude) requires longer fiber length
and higher ratio of L/W to support taller culmthan those growing
at higher latitude (altitude).

Table 5. Correlation coefficients between bamboo wood properties and environmental factors
Index Longitude Latitude Altitude Annual mean
temp.
Minimum
temp.
Annual
precipitation
Relative
humidity
Frostless
days
January monthly
mean temp.
July monthly
mean temp.
1%NaOH extractive -0.465 -0.472 -0.160 0.273 0.320 -0.496 0.686+ 0.323 -0.178 -0.241
Benzo-alcohol extractive -0.437 0.062 0.246 -0.285 0.067 -0.144 0.026 -0.084 0.565 -0.767*
Fibrin -0.106 0.302 0.224 -0.372 -0.437 -0.052 -0.066 -0.784* 0.552 -0.232
Lignin 0.240 -0.087 -0.301 0.035 0.308 0.322 -0.195 -0.164 -0.06 0.181
Fiber Length -0.234 -0.540 -0.675+ 0.355 0.220 -0.614 0.63+ 0.243 -0.555 0.306
Fiber Width -0.188 0.563 0.218 -0.552 -0.284 0.065 -0.863** -0.138 0.447 -0.313
Ratio Of L/W -0.063 -0.631+ -0.492 0.508 0.318 -0.376 0.823* 0.196 -0.531 0.323
Notes: **,*,+ indicate statistically significant at 0.01 , 0.05 and 0.1 level respectively.

Three fiber dimension indexes (fiber length, width and ratio of
L/W) had a close correlation with relative humidity (RH), i.e.
there were significant positive correlations (α=0.05 and α=0.1)
between fiber length and ratio of L/W with RH, and there were
Journal of Forestry Research (2009) 20(3): 261−267

265
very significant negative correlation (α=0.01) between fiber
width and RH. The reason may be that high RH favors the
growth and length of bamboo inter-nodes. Xing (2003) found
that there was a negative significant (α=0.01) correlation be-
tween the number of node under the first branch and RH in Den-
drocalamus latiflorus, which implied that less nodes developed
with higher RH and that these conditions resulted in higher
length of inter-nodes and increase fiber length and ratio of L/W.
There is no obvious geographic variation model in timber
chemistry compositions, although there are some correlations
between chemistry traits and climate factors. There is a signifi-
cant negative (α=0.05) correlation between fibrin content and
frostless days (FD). This result was consistent with the usual
finding that slow-growing tree species have higher fibrin content
compared with that of fast growing tree species. There was also
negative significant (α=0.05) correlation between benzo-alcohol
extract and July monthly mean temperature (Jul. MMT). The
result suggests that higher temperatures might inhibit the growth
of plants and organic compound constructive metabolism.
There were no significant correlations between fiber dimen-
sions, timber chemistry composition and longitude, coinciding
with the findings of similar research on Pinus massoniana by
Zhou et al. (1995).

Genetic manipulation of timber characteristics for different traits

The broad-sense heritability of benzo-alcohol extract and fiber
width could not be estimated because the variance due to prove-
nance of the traits was less than that due to the environment.
Heritabilities for other five bamboo wood traits, such as 1%
NaOH extract, fibrin content, lignin content, fiber length and
ratio of L/W were 0.54, 0.71, 0.38, 0.84 and 0.13, respectively.
Obviously, most of the bamboo wood properties in this study
were under medium or intense genetic control except the ratio of
L/W, benzo-alcohol extract and fiber width. The results suggest
that it would be effective to carry out direct selection for those
high heritability traits in order to improve production and quality.
Selection efficiency for bamboo wood features will be im-
proved by the indirect choice for growth trait or culm form trait
if there are close correlations between the timber features and
growth trait or culm form traits. Correlation analysis for growth
amount and culm form traits (including DBH, height, clear
height, node number under the first branch, culm number per
clump) (Table 6) showed that the chemistry characteristics were
independent from clear height and node number under the first
branch. The 1% NaOH extract was significantly negatively cor-
related with culm number per clump (α=0.05). The correlation
between fibrin content and DBH was significantly positive
(α=0.1). There was no close correlation between lignin content
and growth traits. The results show that these relationships be-
tween bamboo wood and growth or culm form traits are useful
information for pulp-bamboo selection because of the decrease
of extraction with the increase of bamboo output.
Stronger relationships between fiber dimensions and culm
form traits were found in this study (Table 6). For example, there
were significant (α=0.05) or very significant (α=0.01) negative
correlations between fiber width and growth traits (DBH and
height) or culm form index (clear height and node number under
the first branch), and there was very significant negative correla-
tions between fiber length and culm number per clump, signifi-
cant (α=0.05) or very significant (α=0.01) positive correlations
between L/M ratio and bamboo growth or culm form traits.

Table 6. Correlation coefficient between bamboo wood trait, growth trait and culm of Bambusa chungii
Index 1 2 3 4 5 6 7 8 9 10 11
1%NaOH extractive 1 1.000
Benzo-alcohol extractive 2 0.247 1.000
Fibrin 3 -0.331 0.065 1.000
Lignin 4 -0.539** -0.155 0.359+ 1.000
Fiber length 5 0.408 -0.101 0.047 0.060 1.000
Fiber width 6 -0.047 0.071 -0.015 -0.203 0.036 1.000
L/W 7 0.189 -0.095 0.031 0.248 0.354+ -0.915** 1.000
DBH 8 -0.156 -0.219 0.346+ 0.187 0.016 -0.620** 0.572** 1.000
Height 9 -0.130 0.030 0.236 0.028 -0.174 -0.608** 0.480* 0.803** 1.000
Culm 10 -0.512* -0.340 -0.148 0.290 -0.635** -0.075 -0.189 -0.069 0.001 1.000
Clear heigh 11 -0.211 -0.119 0.072 0.164 0.002 -0.536** 0.483* 0.740** 0.626** -0.039 1.000
Node number 12 0.148 -0.258 0.025 -0.203 0.094 -0.431* 0.435* 0.658** 0.621** -0.194 0.641**
Notes: **,*,+ indicate statistically significant at 0.01, 0.05 and 0.1 level respectively.

Selection for optimal provenances of pulp-bamboo

Principal coordinate analysis was carried out with some of bam-
boo wood properties, main growth characters and yield feature
with significant variance difference among eight provenances.
The cumulative proportion of the first, second and third principal
components exceeded 85%, what gives strength to the conclu-
sions of this analysis (Table 7). The first principal component
represents synthesis information of DBH, culm number per
clump, clear height and 1% NaOH extract. The second principal
component represents mainly the information of bamboo wood
yield per clump and fibrin content, while the third principal
component reflects fiber length and fibrin content contrast to
height. The three-dimensional scatter plots of 8 Bambusa chungii
provenances based on the first, second and third axis of principal
coordinate analysis can be used to assess the quality of prove-
nances for good pulp-bamboo (Fig. 2). The results suggest that
Huaiji provenance is an optimum pulp-bamboo provenance with
Journal of Forestry Research (2009) 20(3): 261−267

266
high bamboo wood yield and fibrin content, growth amount and
the lowest 1% NaOH extract. Guilin and Xinyi were the other
two good provenances for developing pulp-bamboo stand bases.

Table 7. Eigenvector of selected principal components
Items
1%NaOH
extractive
Fibrin Fiber
length
DBH Height Culm number
per clump
Clear
height
Node
number
Yield per
clump
Proportion
(%)
Cumulative
(%)
Prin1 0.3428 0.1634 0.3089 0.4321 0.3096 -0.4638 0.3362 0.3869 0.0022 47.3
Prin2 -0.3393 0.5232 -0.2981 0.2083 0.2849 0.0902 -0.0001 -0.0385 0.623 26.9 74.27
Prin3 0.1712 -0.4141 -0.5867 -0.1154 0.4893 0.2054 0.1037 0.3853 -0.0308 10.98 85.25


Fig. 2 Scatter plots of 8 Bambusa chungii provenances based on
first, second and third components of principal coordinate analysis

Discussion

Bamboo has already been used for papermaking in China for a
long time; for instance, Xuan paper (a representative bamboo
product) emerged in the Chin Dynasty. Previous studies showed
that the performance of bamboo paper is superior to that from
grasses, and just similar to that from broad-leaf trees, although it
is inferior to that from conifers (Yang et al. 2002). According to
this study, B. chungii could belong to top-class resource material
for paper-making in bamboo species with features of fiber length
more than 4 mm, ratio of L/W over 300%, HNO3-alcohol fibrin
content exceeding 44%, lignin content less than 26% (Hui et al.
1996)
Zobel (1971) found wood density, branch diameter and stem
straightness were affected by genetic factors in loblolly pine
(Pinus taeda). Also, families with characteristics of fast-growth
and high-density have been selected based on the knowledge of
Zobel. Extensive variance in wood properties was found in pop-
lar (Jiang et al. 1994), spruce (Yanchuk et al., 1992), P. masson-
iana (Zhou et al. 1995), Larix kaempferi (Sun et al. 2003), Cun-
ninghamia lanceolata (Sun et al. 1993). So, the utilization of the
genetic variance is one of the main tasks of forestry genetic im-
provement. However, there are few study on intraspecies geo-
graphic variation of bamboo wood properties. In this study, some
chemical composition variances were found for the first time
among B. chungii provenances, optimum provenances with de-
sired features could be selected, and some chemistry traits varied
markedly among individuals but not among provenances. For
those traits the selection of individual plants can be a good
choice. The selected individual can be extensively utilized by
large-scale cloning propagation.
Geographical origin is sometimes one of the best indicators of
genetic performance because the difference occurs in chemical
composition of wood in different geographical origin. Take Oak
Barrels as example, the species from United States and France
can be easily distinguished from each other by the detection of
chemical composition of wood. Most of the species from France
can be unambiguously assigned to their actual classes of geo-
graphical origin even though they were collected from neighbor-
ing forests (MARCO et al. 1994). In this study, fiber length and
ratio of L/W in B. chungii have decreasing geographical trends
from south (low altitude) to north (high altitude) in China. Our
study result is agreed to previous studies (Wang et al. 2007; Har-
ris 1965). Generally, there was no significant correlation between
the fiber morphology and longitude. Piedra et al. (1986) found
that the differences in tracheid length had no significant changes
in mature P. tecunumanii from east to west. For latitude, B.
chungii behaves in a similar way as P. tecunumanii. So, the con-
clusion suggests that the researchers should pay more attention to
changes of fiber characters when introducing B. chungii from
south to north to ensure that the new plantation keeps its genetic
superiority.
Wood properties are assumed to be under the control of mul-
tigene families. As a result of that, there are different degree of
correlations among wood properties and other traits. When con-
sidering wood properties for tree improvement programs, the
interaction among them and with other traits should be investi-
gated. Correlation coefficients between microfibril angle and
wood properties showed that they have a significant negative
relation in poplars (α=0.01; Fang et al. 2004). There was a strong
or moderate negative genetic relationship between diameter
growth rate and wood density in Picea glauca (Carriveau et al.
1987) and others. In our study, relationship between bamboo
traits and properties for B. chungii was investigated for the first
time, providing useful information for the selection of genetic
materials for paper-making.

Acknowledgments
The authors are grateful to Forestry Bureau of Maoming City for
its technical support in the field, Prof. Manuel Perez Ruiz for his
kind help in English writing and Prof. Liu Li for her assistance of
timber chemical components analysis.
Journal of Forestry Research (2009) 20(3): 261−267

267

References

Carriveau A, Beaulieu J, Mothe F. 1987. Wood density of natural white spruce
populations in Quebec. Can J For Res, 17: 675–682.
Fang Shengzuo, Yang Wenhong, Fu Xiangxiang. 2004. Variation of microfi-
bril angle and its correlation to wood properties in poplars. Journal of For-
estry Research, 15(4): 261−267.
Gan Jiaqi. 2002. Prospects of bamboo stands development and utilization for
pulping and papermaking. Southwest Pulp and Paper, 2: 8−10. (in Chinese)
Harris JM. 1965. A survey of the wood density, tracheid length, and latewood
characteristics of radiata pine grown in New Zealand. New Zealand Forest
Service Forest Research Institute. Technical paper, 4: 34.
Hui Chaomao, Du Fan, Yang Yuming. 1996. Bamboo Cultivature and utiliza-
tion. Beijing: China Forestry Publishing House, pp.123. (in Chinese)
Hui Chaomao, Wang WenJiu. 1993. Studies on the bamboo resources from
shuanjiang county and the bamboo wood property for pulping. Journal of
Southwest Forestry College, 13(2): 79−87. (in Chinese)
Jiang Xiaomei, Zhang Lifei, Zhang Qiwen, et al. 1994.Genetic Variation in
Basic Wood Properties of 36 Clones of Populus deltoids. Forest Research,
7(3): 253−257. (in Chinese)
Ma Lingfei, Zhu Liqing. 1990. Fiber Forms and Tissue Percentage of Six
Species of Sympodial Bamboos in Zhejiang Province. Journal of Zhejiang
Forestry College, 7(1): 63−68. (in Chinese)
Ma Naixun, Zhang Wenyan, Chen Guangcai. 2004. Suggests about accelerat-
ing development of using bamboo timber for paperming in China. China
Forestry Science and Technology, 18(1): 9−11. (in Chinese)
Marco J, Artajona J, LARRECHI MS, Rius FX. 1994. Relationship Between
Geographical Origin and Chemical Composition of Wood for Oak Barrels.
Am J Enol Vitic, 45(2): 192−200.
Matheson AC, Turner CH, Dean GH. 1986.Genetic variation in the pulp
qualities of Eucalyptus obliqua L. Herit Appita, 39(3): 205−212.
Piedr, TE, Zobel BJ. 1986.Geographic variation in wood properties of Pinus
tecunumanii. Wood Fiber Sci, 18: 68–75.
Qin G.uofeng. 2003. Geographic provenances of Pinus massoniana. Hang-
zhou: Zhejiang University Press, p189. (in Chinese)
Sun Chengzhi, Xie Guoen, Li Ping. 1993.Variations of wood properties and
characteristics and a tentative selection for structural timbers of cunning-
hamia lanceolata provenances. Scientia Silvae Sinicae, 29(5):429−437. (in
Chinese)
Sun Xiaomei, Zhang Shougong, Qi Liwang, Wang Junhui, Lu Shoufang, Jiang
Yingshu. 2003. Genetic Variations in Pulpwood Qualities of
Open-pollinated Japanese Larch Families. Forest Research, 16(5): 515−522.
(in Chinese)
Wang Qiuyu, Qu Lina, Jia Hongbai. 2007. Variation of Wood Fiber Charac-
teristics, Microfibril Angle and Basic Density of Betula platyphylla in
Natural Populations. Journal of Northeast Forestry University, 35 (2): 1−3,
6. (in Chinese)
Wang Wenjiu, Hui Chaomao, Liu Cui, Wang Changming, Chen Yuhui, Fu Hui.
1999. A study on the chemical compositions of 14 timber bamaoo species
in Yunnan Province. Journal of Bamboo Research, 18（2）：74−78. (in
Chinese)
Wang Wenjiu, Hui Chaomao, Liu Cui, et al. 1999. A study on the chemical
compositions of 14 timber bamaoo species in Yunnan Province. Journal of
Bamboo Research, 18(2): 74−78. (in Chinese)
Wright JA, Baylis B. 1993.Volume, pulp and papermaking traits of Pinus
maximinoi provenance planted at two sites in South Africa. South Africa
Forestry Journal, 165: 37−40.
Xing Xinting. 2003. Study on the genetic variation of populations and im-
proved seeds breeding of Dendrocalamus latiflorus Munro. Dissertation for
the PhD Degree of Chinese Academy of Forestry. Beijing: Chinese Acad-
emy of Forestry, p30, 35. (in Chinese).
Yanchuk AD, KISS, GK. 1992. Genetic varation of growth and wood specific
gravity and its utility in the improvement of interior spruce in British Co-
lumbia. Silvea Genetica, 42: 141−148.
Yang Rendang, Chen Kefu. 2002. Performance and Potential of Bamboo as
Papermaking Material. China Forest Products Industry, 29(3): 8−12. (in
Chinese)
Zhou Zhichun, Qin Guofeng, Li Guangrong, Huang Guanglin, Chen Bingxing,
Cheng Chuanyan. 1995. Geographic Pattern of Chemical Components in
Wood and Pulp/Paper-making Properties in Natura1 Stands of Masson Pine.
Forest Research, 8(1): 1−6. (in Chinese)
Zobel B. 1971. Genetic manipulation of wood of the southern pines including
chemical characteristics. Wood Science and Technology, 5(4): 255−271.