Taste analyzers, developed in Japan, convert various physicochemical p
arameters of rice into ''taste'' scores based on correlations between
near-infrared reflectance (NIR) measurements of key constituents (e.g.
, amylose, protein, moisture, fat acidity) and preference sensory scor
es. These taste analyzers are being used by Japanese millers and whole
salers to grade rice, both domestic and imported. This study examines
the applicability of using the Satake Neuro Fuzzy Rice Taster for eval
uating U.S. medium-grain rice cultivars. The Rice Taster, as presently
calibrated, does not appear to be a valid tool for assessing rice cul
tivars with low (<18%) amylose contents. The low-amylose cultivars fel
l outside the range of the calibration set used by Satake. The effects
of degree-of-milling and U.S. shipping practices on Rice Taster score
s (S) were also determined. Conditioning of the bran during four weeks
refrigerated storage led to the bran being more readily removed from
the kernel during milling. This resulted in the milled rice having sig
nificantly lower protein (P < 0.0001), free fatty acids (P < 0.01), an
d n-hexanal (P < 0.0001) levels, increased whiteness (P < 0.0001) and
milling degree (P < 0.0001) measures, and higher S (P < 0.0001) values
. Deep milling significantly increased (P < 0.0001) chemical measures
of amylose and significantly decreased protein and free fatty acids co
ntents (P < 0.0001). Rice Taster measurements of ''amylose'' (A) and p
rotein (B) significantly decreased (P < 0.0001) with deep milling. Moi
sture (C), ''fat acidity'' (D-a), and milling yield (D-b) were not sig
nificantly (P > 0.05) affected. S significantly (P < 0.0001) increased
(mean 5 points) with deep milling.