### Table 3: Curvature measurements.

"... In PAGE 5: ... The tangents apos; rate of change provided an estimate of local curvature at the contact. Table3 shows the mean and standard... ..."

### Table 1 Curvature classes

in Corner

1998

"... In PAGE 5: ... As a result, H gxxgyy g2 xy 10 and K gxx gyy 2 : 11 The signs and zeros of the mean and Gaussian curvatures can be used to categorise the local surface geometry into a number of distinct topo- graphic classes. These classes are summarised in Table1 . In this paper, we are interested in the saddle-structures which are labelled as hyperbolic features in the table.... ..."

### Table 1: Curvature classes

"... In PAGE 5: ... According to the definitions given above K = 1 2( + ) (8) H = ? 2 (9) The signs and zeros of these two quantities can be used to label the surface according to curvature class. The different classes are defined in Table1 . In Figure Figure 2 we show the relationship between the classes.... In PAGE 10: ... Specifically, P(n)(Si = !) is the weight of evidence assigned to the label assignment ! at site Si for the iterative epoch n of the algorithm. These weights are initialised with the probability that label ! takes on one of the eight possibilities from Table1 . Initially, these probabilities are calculated using the computed values of H and K, together with their known covariance structure.... ..."

### Table 3 Results of a t-test for measured whole deviations and simulated whole deviations. (r: correlation coefficient; t: t values; df: degrees of freedom; *p , 0.05; **p , 0.01; HJ: minimum hand jerk criterion; AJ: minimum angle jerk criterion; TC: minimum torque change criterion; CTC: minimum commanded torque change criterion)

"... In PAGE 9: ... In previous studies, Hollerbach (1990); Osu, Uno, Koike, and Kawato (1997) suggested that planning in the joint space cannot explain a gently curved hand trajectory. As shown in Table3 , the slopes relating whole deviations of the minimum hand jerk and actual trajectories were zero (the mean error the slope was 1 in both planes). For mini- mum angle jerk trajectories, slopes were from 1.... In PAGE 9: ... Moreover, the minimum commanded torque change criterion most similarly reproduced the feature of the measured trajectory in point- to-point movements on the sagittal plane. The whole deviations in minimum commanded torque change trajec- tories on the horizontal plane and the sagittal plane shown in Table3 predict that the curvatures of the trajectories on both planes have almost the same tendency. Therefore, results such as a linear relation between the whole deviation of a trajectory measured in a horizontal plane and the whole deviation of a trajectory measured in a sagittal plane seem appropriate for the minimum commanded torque change criterion.... In PAGE 12: ... As a result of the analysis, we plotted the correlations between the whole deviations of measured and predicted trajectories. Table3 summarizes the correla- tion coefficients, the results of a test on the correlations, and the slopes of the regression lines of all of the subjects in the horizontal and sagittal planes. All of the minimum hand jerk trajectories were straight; this indicated that they were not correlated with those of actual trajectories.... ..."