Air-dried Japanese cypress (Chamaecyparis obtusa Endl) was used as the test specimen. The test specimen dimensions were 390 mm (longitudinal) × 80 mm (radial) × 80 mm (tangential). Moisture content and density were 7.9% and 0.45 g/cm³, respectively.

A surface SH-wave was generated on the surface of a Japanese cypress column as shown in Figure 1. Wedge transducers made a shear wave incident obliquely with respect to the lumber surface. The angle of incidence was 48.59˚. The wedges were made from polyetherimide resin. The measurement of receiving waveforms was performed by using an ultrasonic pulser-receiver (models JPR-10CK by Japan probe, Japan) and shear wave transducers with a natural frequency of 500 kHz and diameter of 1 inch (models CR-0016-SA by Staveley Instruments, USA), as shown in Figure 1. In addition, epoxy resin (SH111) was used as the coupling medium to improve the bonding between the transducers, the wood specimen and the wedges.

The wedge transmitter was fixed at the wood specimen edge. The wedge receiver was slid along the wood longitudinal direction, as shown in Figure 2. The receiving waveforms of surface SH-wave were recorded on a personal computer every 50 mm. The measurement positions were the four sides of the wood specimen (A, B, C, D), three lines per one side (L₁, L₂, L₃).

Propagation time was calculated by using the peak amplitude of the first receiving waveform with zero crossing method. The propagation time measurement was repeated three times, and the resulting experimental values were averaged.

The surface SH-wave velocity was calculated from Equation (1):

where T is the propagation time and n is the receiver’s position.

Following the ultrasonic measurements, wood specimen was cut into small pieces, each 10-mm thick. These pieces were used for measuring air-dried density and moisture content. Air-dried density was measured using the water displacement method. Sliced specimen pieces were dried overnight in a constant-temperature oven at 105˚C. The moisture content of wood was calculated as the ratio of the water content weight to the oven-dried weight.

Figure 3 shows the typical receiving waveforms of the surface SH-wave. The inter-wedge distances were 100, 150, 200, 250, and 300 mm. As the inter-transducer distance increased, the receiving waveforms moved to the right. Figure 4 shows the propagation time T_R and the receiving wave amplitude A_R vs. the measurement position. T_R and A_R were obtained from the waveform in Figure 3. As shown in Figure 4, T_R increased from 130 µs to 270 µs as the wedge distance L_W increased. T_R was positively correlated with L_W at 1% of significance level (r = 1.00). On the other hand, A_R decreased from 2.0 mV to 0.6 mV as the wedge distance L_W increased. A_R was negatively correlated with L_W at 1% of significance level (r = 0.99).

When the knots were located between the wedges, the amplitude of the receiving waves became smaller or non-detectable. Sandoz reported that ultrasonic wave velocities in the longitudinal direction were negatively

correlated (r = −0.48) with the number of knots on the spruce lumber surface [16] . Additional experiments are needed to determine the effect of knot size on the propagation characteristics of SH-waves.

Figure 5 shows the distribution of surface SH-wave velocity. Velocity ranged from 1270 m/s to 1496 m/s. The averaged velocity was 1384 ± 57 m/s. Bucur et al. reported surface wave velocity for spruce timber ranging from 1113 m/s to 1210 m/s [17] . The experimental values in this study almost agreed with those obtained in the previous studies. The value of velocity varied depending on the measurement location. The averaged velocities in the outer part (L₁, L₃) and the central part (L₂) were 1406 ± 52 m/s and 1342 ± 44 m/s, respectively. As shown in Figure 5, the velocities in the central part were slower than those in the outer part. Wood is a biomaterial with three orthotropic axes. Specifically, for softwood such as Japanese cypress more than 90% is contributed by the axial cells (i.e., tracheids), which are highly elongated in the grain direction. Tracheid structures determine the physical and mechanical properties of wood. An ultrasonic wave velocity is different in the propagation direction and polarization direction [17] - [20] . Bucur et al. concluded that for softwood and hardwood the longitudinal and shear wave velocities depend on the three orthotropic directions [17] . For Japanese cypress, Hasegawa et al. reported the longitudinal wave velocities in the three directions as 4492 m/s, 1865 m/s, and 1389 m/s [18] . In addition, Mishiro reported the longitudinal wave velocities in the three directions as 5010 m/s, 2050 m/s, and 1220 m/s [19] . Bucur reported the shear wave velocity polarized into the radial and tangential directions in wood as 1322 m/s and 1229 m/s for spruce (Picea abies (L.) Karst) and as 1517 m/s and 1273 m/s for beech (Fagus sylvatica L) [20] . In the present study, the SH-wave polarization direction corresponded to the tangential direction in the central part, and to the radial direction in the outer part. For these reasons, the velocities in wood were different for different locations in the same symmetry plane.

Figure 6 shows the relationship between the velocity and the moisture content. Velocity and moisture content were negatively correlated at 1% of significance level (r = 0.38). Longitudinal wave velocity in the longitudinal direction decreased with increasing moisture content until the fiber saturation point (FSP) was reached [21] [22] . In this study, the moisture content was 6.4% to 8.7% below FSP. The dependence of velocity on the moisture content was the same as that observed in the previous study. However, few studies have reported the relationship between shear wave velocity and moisture content. In this study, both dependences were experimentally confirmed for the first time. These results suggest that the moisture content effect has to be considered for applying surface SH-wave acoustoelastic technique to drying stress measurement in wood.