Based in part on observations collected at the Observatoire de Haute-Provence.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

Tables 9-11 plus detailed tables for each object are available in electronic form at the CDS via anonymous ftp to cdsarc.u-strasbg.fr (130.79.128.5) or via http://cdsweb.u-strasbg.fr/cgi-bin/qcat?J/A+A/383/30

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... success

This assertion is questionable as pointed out by the referee. We therefore tested it by comparing the distributions of the gradients in subsamples of 15 to 25 objects, sorted out by NGC numbers, from the surveys of Peletier et al. (1990), Goudfrooij et al. (1994) and our discussion in RM00. This procures 8 tests, according to the number of colours in each of the sources. It turns out that the subsamples are statistically coincident in 4 cases. They differ by slightly more than the calculated errors in the others.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... data

The coefficients of correlation given here are of course different from the previously calculated ones in Sect. 3.2.7.6, which refered to other colours with smaller gradients.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.