Table II.
ζlow | λ | R2 | λ1 [cm−1] | R2 | ||||
---|---|---|---|---|---|---|---|---|
0 | 0.978 | 0.909 | −0.0825 | 0.981 | ||||
0.1 | 0.974 | 0.939 | −0.0838 | 0.983 | ||||
0.2 | 0.947 | 0.986 | −0.0917 | 0.986 | ||||
0.3 | 0.904 | 0.993 | −0.0917 | 0.976 | ||||
0.4 | 0.883 | 0.995 | −0.1009 | 0.985 |
For each value of ζlow the average branching rate λ along the “main path” of each tree was computed from the slope of a linear regression of log(r n/ r 0) versus n, cf. Eq. 11. These regressions against the structural variable n were seen to discriminate well between trees. Conversely, regressions for r n/r 0 over the metric path lengths (regression slope λ1) failed to discriminate between trees. R2 is the ratio of explained variance (i.e., R2 = Sum of Squares (Model)/Sum of Squares (total), used as the criterion for the goodness of fit.