Part of the observed HI-profiles given in Fig. 5 were recognized as confused, because one or several other galaxies are within the beam. These objects are marked with an "a'' or a "c'' in the last column of Table 2, and individual comments, mainly about possible companions or data quality, may be found in Appendix A; references related to these comments are listed in Appendix B.
Considering possible selection effects, no dependence of the
detection rate on inclination, line width, or distance,
were observed in the investigated redshift range
( km s-1).
However, the 10 percents of missing objects (i.e. unsuccessfully observed)
indicates a possible cut-off of the
distribution.
It is indeed probable
that, due to noise fluctuations or base-line effects, one tends to miss
very narrow as well as very wide HI-profiles. This means that one
has to be careful when determining for example the Hubble constant H0
from the inverse Tully-Fisher relation, because a systematic effect,
comparable to the Malmquist bias in the direct case, is expected.
This debate is presented in details by Ekholm & Teerikorpi (1997),
and will be further discussed in a forthcoming paper.
It has been shown also that the HI detection rate is slightly correlated
with morphological type. This (expected) result implies that it is
necessary to process separately the different types
when using either the inverse TF relation, or the TF-21
relation (relation between the 21-cm magnitude and ).
We have made use of data from the Lyon-Meudon Extragalactic Database (LEDA) compiled by the LEDA team at the CRAL-Observatoire de Lyon (France).
Copyright The European Southern Observatory (ESO)