The panelists were trained in the descriptors evaluation using
commercial products with different texture characteristics. It was
performed six training sessions of 50 min each one. The first
session consisted in a meeting where the panel leader explained
the use of the 10-cm unstructured scale anchored with the
intensity terms at the endings, for example, ‘‘low or none’’ at the
left and ‘‘very’’ at the right. In the second session, the intensities of
the texture attributes were scored. All samples were evaluated in
each session. Sessions missing were conducted until the panel was
homogeneous in its assessments and its comprehension of the
extremes of the scales. The samples, cut in cylinders of 1.8 cm
diameter and 1.5 cm height, were served at 20 8C on plastic
containers, labeled with three-digit random numbers. Mineral
water was used as rinsing between samples. Evaluations were
carried out in the sensory laboratory of LAN/ESALQ/USP, under
artificial daylight type illumination, temperature control
(20 -
2 8C) and air circulation. The panelists were selected using
four criteria (discrimination, reproducibility, agreement and use of
the scale) (Stone & Sidel, 2004) based on the panel performance