10 another. Using a needle dropper, the macerated fibre suspension was placed on a
standard slide of 7.5×2.5 cm based on TAPPI T401 om-88 procedure TAPPI 1991. Thirty slides of macerated fibres were prepared from each segmented ring
segment of 1 cm width. The slides were then dried and a cover glass of 22×30 mm was placed over the fibres. Fibre length was measured under an optical video
microscope. Undamaged single fibre was selected from each slide and its image was taken. The captured images were analyzed using Motic image software for
measurement of the fibre length. The number of measured fibers were 30 fibers in each segmented ring.
Using a microtome, thin sliced specimens were prepared for measurement of microfibril angle. Juvenility test specimens Figure 3-A3 inserted into the
microtome holder and sliced to produce undamaged thin slices of 30 μm thick. The undamaged thin slice was transferred onto a slide of 7.5×2.5 cm which had a
few drops of distilled water using drawing brush. Fifteen slides were prepared from each segmented ring. The slides were then dried and a cover glass of 22×30
mm was placed over each thin slice specimen. The slides were analysed under a light microscope to find cells containing microfibrils. Images of the cells were
captured and analyzed using Motic image software for measurement of the microfibril angle.
2.3.8 Transition age
A segmented regression analysis was used to determine the demarcation point from juvenile to mature wood. It was assumed that development of density,
fibre length and microfibril angle from pith to bark can be described by two functions in a curve. The first function was a steep slope of the curve over the first
few years beginning at the pith juvenile wood and the second function was a constant slope for the later part of the curve mature wood. The fitted regression
model for the functions took the form of quadratic model with plateau Figure 5. The change of slope in the radial fibre length, microfibril angel and density trends
as a function of segmented ring number was modelled as follows Equation 2:
Yi = A + BXi + CXi
2
+ Ei……… 2 where Yi = dependent variable fibre length, microfibril angle, density, Xi =
segmented ring number, A = intercept of the line of the juvenile wood, B and C = regression coefficients and Ei = error.
Figure 5 Fitting a segmented model using NLIN non linear Where, y = dependent variable, a = intercept, b and c = regression
coefficients for the equation, x = experimental data Quadratic
y = a + b x + c x
2
Plateau y = p
x X
Y
11 From theoretical considerations, it can be hypothesized that:
y = a + bx + cx
2
if x x , the equation relating y and x is quadratic and
y = p if x ≥ x , the equation is constant horizontal line
where x = ring number at which wood changes from juvenile to mature wood, p
= fibre length or microfibril angle or density at which wood changes from juvenile to mature wood, a = intercept, b and c = regression coefficients for the equation.
With segmented regression, the statistical model Equation 2 can simultaneously estimate parameters of the model and a demarcation point between
juvenile and mature wood. The demarcation point can be directly obtained using non-linear least squares procedures PROC NLIN in SAS 1990, Version 6,
which minimizes the mean squared error. The PROC NLIN in SAS was used to obtain estimates of regression parameters and the demarcation point. PROC NLIN
could fit segmented model as in Figure 5. The curve in Figure 5 must be continuous the two sections must meet at x
and the curve must be smooth the first derivate with respect to x are the same at x
. These conditions implied that x = -b2c, and p = a
– 2b4c where b and c = regression coefficients, p = fibre length or microfibril angle or density, at which wood changes from juvenile to
mature wood.
2.4 Results and Discussion
2.4.1 Density
Average density of 5, 6 and 7 years old sengon and jabon wood, 26 years old douglas-fir and 18 years old poplar cultivars are shown in Figure 6a-d. Generally,
densities of sengon, jabon and douglas-fir tended to increase from pith to bark. Poplar cultivars
‘lambro’, ‘soligo’, ‘koster’ and ‘I214’ showed different trend. It was due to the samples were taken from veneer samples not solid wood samples.
Therefore, the wood near pith was not represented enough in poplar samples. Douglas-fir had the highest density in average of 726 kgm
-3
, followed by jabon ’s
density 437 kgm
-3
, poplar cultivars’ density 401 kgm
-3
and sengon was the lowest 331 kgm
-3
. Density of 5 years old, 6 years old and 7 years old sengon wood near the pith
were 237 kgm
-3
, 259 kgm
-3
and 248 kgm
-3
, respectively Figure 6a. Densities of sengon wood near the bark were 393 kgm
-3
5 years old, 456 kgm
-3
6 years old and 442 kgm
-3
7 years old. Martawijaya et al. 2005 stated that the density of sengon wood ranged from 240 to 490 kgm
-3
average 330 kgm
-3
. Wood density near the pith for the 5 years old, 6 years old and 7 years old
jabon wood were 234 kgm
-3
, 297 kgm
-3
and 284 kgm
-3
, subsequently Figure 6b. While the densities of jabon wood near the bark were 573 kgm
-3
5 years old, 606 kgm
-3
6 years old and 615 kgm
-3
7 years old. Martawijaya et al. 2005 reported that the density of jabon wood, 290 to 560 kgm
-3
average 420 kgm
-3
, although no information was given on whether the samples were near the bark or
pith. For 26 years old douglas-fir, wood density near the pith was 405 kgm
-3
, and wood density near bark was 570 kgm
-3
Figure 6c. Martin et al. 2006 reported that the density of 31 years old Douglas-fir is 447 kgm
-3
, while CIRAD 2011 found solid Douglas-fir density is 540 kgm
-3
although no information was available whether the samples were taken from wood near the bark or pith.
