Previous research indicates that perceived orientation and/or alignment of
segments and points san vary as a function of the angular position of the s
timulus elements. Several studies show that the variability of the response
s is least and accuracy of judgment is greatest where segments and dots are
aligned with a cardinal axis. Additionally, some report assimilation of ju
dgments toward the nearest cardinal axis-that is, the segments (or dots) ar
e seen as being closer to the horizontal or vertical than is true. The pres
ent research confirms that judgments of collinearity are least variable and
most accurate when the segment being judged is aligned with a cardinal axi
s. However, we do not find any consistent tendency for cardinal axis assimi
lation. Plotting the collinearity error (delta) as a function of angular po
sition (phi), we find a distinctive profile of oscillation for each subject
. Furthermore, subjects who were evaluated in two sessions showed very simi
lar profiles of delta oscillation from Day 1 to Day 2. Harmonic analysis in
dicated a wide-ranging pattern of significant components. The components at
the 4th harmonic and below were more likely to be significant, but each su
bject showed differential loadings in terms of which of the components were
significant, as well as in the sign and amplitude of significant component
s. These results may reflect idiosyncratic fixation tendencies, or individu
al differences in the design of neural mechanisms that encode the angular p
ositions of stimuli.