improvement in rice quality. CV of grain number per panicle
was lower at FPM (13.4%) than at VRF (9.1%). The low CV of
grain number per panicle seemed to be related with low CV of
panicle numbers (Table 4). Although there was no significant
difference in rice yield between VRF and FPM, the CV of rice
yield at VRF (5.8%) was much lower than FPM fields (8.1%).
Rice quality characteristics measured in brown, milled, and
cooked rice are shown in Table 6. Head rice ratio of brown rice
was not different between VRF and FPM, but head rice ratio of
milled rice was higher than that of brown rice. Head rice ratio of
milled rice was 94.7% at VRF being significantly higher than
88.4% at FPM. Head rice ratio and its difference between VRF
and FPM increased by the milling process. On the other hand,
protein content of brown rice at VRF was 6.9%, which was sig-
nificantly lower than 7.1% at FPM. Nitrogen top-dressing rate at
panicle initiation has a significant effect on grain yield and the
protein content of milled rice (Lee et al. 2009). The low protein
content of brown rice in VRF appears to be as a result of
reduced amount of applied N fertilizer. However, the difference
in protein content was not evident in milled rice. Seed protein is
highly distributed in aleurone layer, and the amount of soluble
protein in rice endosperm has a significant effect on the texture
of cooked rice (Tamaki et al. 1989; Zakaria et al. 2000). There
was no significant difference in the protein content of milled rice
between FPM and VRF. The lower protein content of milled
rice suggests that most of protein in aleurone layer was removed
during the milling process and this process have a direct influ-
ence on appearance and ingredient quality. Head rice yield was
higher by 0.26 t ha-1 due to the greatly improved head rice ratio