Because many galaxy properties like color, star formation rate and
emission line strength correlate with magnitude (see, e.g. Moody and
Kirshner 1988), our sample must span a broad range in absolute
magnitude. As a result, we have avoided a purely magnitude or diameter
limited sample. We select galaxies form the first CfA redshift
catalogue, which contains galaxies to a limiting magnitude of
mZw = 14.5 (Huchra et al. 1983).
For observational reasons, we wanted to minimize the number of galaxies subtending an angle on the sky of more than 3 arcmin. We did not want to enforce a strict diameter limit, which would e.g. enhance the bias against lower surface brightness galaxies already present in the CfA catalog (like in any magnitude limited catalog). We chose to adopt a luminosity dependent velocity cut, where intrinsically bright galaxies had to be farther away than intrinsically fainter systems, in order to be selected. Luminosities were calculated asuming a simple linear Hubble flow and H0=100 km/s/Mpc. Also, we had to exclude galaxies in the Virgo Cluster, as inclusion would bias our sample towards a high density cosmic environment. After these cuts, ~1000 of the ~2400 galaxies in the CfA catalog remained.
We sorted the selected galaxies as a function of luminosity in 1 magnitude wide bins of absolute blue photografic magnitude, and sorted the individual bins according to morphological type. We then drew randomly galaxies from each bin in such numbers as to approximately reproduce the observed local luminosity function (e.g Marzke et al. 1994). Our final sample contains galaxies with the desired range of absolute magnitudes (-14 MB -22), extending five magnitudes fainter than the characteristic luminosity MB*~-19.2 (de Lapparent et al. 1989), and fairly sampling the changing mix of morphological types as a function of luminosity.