*Astron. Astrophys. 362, 851-864 (2000)*
### Footnotes
**1.** The FORTRAN codes of the MIST algorithms are available on
request to the authors.
**2.** This representation of the FP is adopted throughout the
paper.
**3.** The OLS method can be thus regarded, in this respect, as
a particular case of the MIST fit, in which the scatter is completely
ascribed to the direction of the dependent variable.
**4.** Notice that the mean errors are often the only
significant available estimates.
**5.** This amounts to a 15% scatter in the
direction, i.e. to the mean value of
the FP scatter of clusters, see Sect. 4.
**6.** The *p*th percentile of a given distribution is the
point that exceeds (100-p) ascending order (see e.g. Beers et al.
1990).
**7.** A percentile interval
is given by the difference between the
th and the
th percentiles (see note 6) of the
distribution of interest.
**8.** I.e. by confidence intervals independent of any
assumption on the shape of the probability density of the statistics
of interest.
© European Southern Observatory (ESO) 2000
Online publication: October 30, 2000
helpdesk.link@springer.de |