Regression of sensory scores on instrument variables could
be used to develop prediction equations to estimate sensory textural
attributes in those cultivars where the ranges of sensory
scores and instrument values were broad. The best single Instron
texture profile variable was similar to Magness-Taylor firmness
in predicting sensory scores; however, the Instron texture profile
variable selected as best differed among cultivars and between
years within cultivars. Using combinations of several Instron
texture profile variables frequently improved prediction (maximum
models in stepwise regression usually contained only 1 or
2 variables but some contained as many as 6 variables). Therefore,
when prediction of the sensory textural characteristics of
apples is important, as in evaluating effects of preharvest or