Sample Selection

  The Canadian Network for Observational Cosmology (CNOC) cluster survey [\protect\astronciteYee et al.1996][\protect\astronciteCarlberg et al.1994] mapped 16 intermediate redshift galaxy clusters selected by X-ray luminosity with the goal of determining cluster mass profiles and the cosmological density parameter . The survey data base contains direct Gunn and MOS images, 2D spectra and extracted 1D spectra for thousands of galaxies. Position, Gunn total magnitude, Gunn color, and redshift were determined by the CNOC collaboration for each galaxy. In order to provide field contamination estimates, 1/3 to 2/3 of the galaxies in a typical CNOC field were field galaxies unassociated with the clusters. This large subset of field galaxies with known properties was ideal for the preparation of efficient multi-object spectroscopy runs.

Four CNOC clusters (EMSS1512, EMSS1621, Abell 2390 and EMSS0015), which were visible at the time the observations were scheduled (July-August 1994), were selected. Using the CNOC spectral classification index Scl assigned to each galaxy, the catalog files went through a first pass to pick out all objects with Scl = 5 (emission line). The Scl index is based on the template giving the best cross-correlation peak. A second pass through the catalog files identified field galaxies. In order for a galaxy to be identified as part of the field, its redshift had to be in the range 0.250.45 and at least 0.02 above or below the mean CNOC cluster redshift. A 11area was extracted around each field emission line object on both the Gunn and images. These postage stamp images were used to calculate galaxy position angles. The position angle of each object was determined in both colors by fitting galaxy isophotes with ellipses using the task ELLIPSE in the IRAF/STSDAS package ISOPHOTE. The position angles were used at the telescope to align slitlets with the major axis of the primary target galaxy (see section ).

The major limiting factor of internal kinematics studies on a 4-m class telescope is flux. [OII] strong objects were selected according to their [OII] equivalent widths. Equivalent widths were measured from the CNOC 1D extracted spectra by fitting a Gaussian to the [OII] line using SPLOT in the IRAF/NOAO package ONEDSPEC. The observed equivalent widths were corrected to their rest-frame value by dividing them by (1+z). Objects with equivalent widths in the range 2050 Åwere selected as targets. This inevitable selection criterion complicates the task of tying intermediate redshift samples with the diverse (and often poorly known) local galaxy population. Although such [OII] strong objects could be locally regarded as extreme, they are representative of the excess galaxy population at intermediate redshifts [\protect\astronciteKennicutt1992][\protect\astronciteBroadhurst et al.1992][\protect\astronciteBroadhurst et al.1988]. Locally, there is a dependence of [OII] equivalent width on galaxy type. The target galaxies may be [OII] strong relative to early-type (Sb and earlier) spirals which have [OII] equivalent width not exceeding 20Å, but they have the same [OII] equivalent widths as many late-type (Sc and later) galaxies [\protect\astronciteKennicutt1992] which have equivalent widths between 20 and 60 Å. The claim made in many studies that local galaxies have [OII] equivalent widths lower than 20 Åis incorrect.

To summarize the target selection process, target galaxies had to be emission-line objects in the field with redshifts in the range 0.250.45 and [OII] equivalent widths greater than 20 Å. No selection was made on the basis of size or color. The next step consisted of maximizing the number of galaxies that could potentially be observed with a multi-object spectrograph. A group was formed around each target galaxy by drawing a square the size of the SIS field (33) with the same position angle as the galaxy's major axis and counting the neighboring galaxies falling within the square. A number of considerations dictated the choice of the best target fields: total number in group, [OII] strength, elongated galaxy images, and a usable guide star for SIS adaptive optics corrections (see section ).

The characteristics of the target galaxies are shown in Table . Column 1 is the galaxy identification (ID). The first part of a galaxy ID is the name of the CNOC cluster around which the field was observed, and the second part is the PPP number identifying a galaxy in the CNOC catalog file. Columns 2-4 are the galaxy redshift, CNOC Gunn magnitude and color. Columns 5-6 are the Johnson B band apparent magnitude and the rest-frame Johnson B absolute magnitude described below. Target galaxies span a wide range of absolute B magnitudes ranging from 21.7 to 18.2. Column 7 is the rest-frame [OII] equivalent widths determined from the CNOC 1D extracted spectra, and column 8 is the expected H rotation velocity corresponding to M according to the absolute calibration of the local B band Tully-Fisher relation (see Section ). Column 9 is the exponential disk scale length measured from fits to the luminosity profiles of the target galaxies (see section ).

Rest-frame absolute B band magnitudes are key quantities in the present study. They were calculated from the CNOC Gunn magnitudes and - colors as follows. M is given by

where is the observed Gunn apparent magnitude, (B) is the color index necessary to convert to B at the redshift of the galaxy, k is the so-called B k-correction, and d is the luminosity distance computed in a H = 75 km s Mpc and q = 0.5 universe. The k-correction is needed to account for the fact that a filter with a fixed bandpass will ``see'' different regions of a galaxy's spectral energy distribution with increasing redshift and that the energy seen by a filter decreases by a factor of (1+z) due to wavelength stretching by the cosmological expansion. The k-correction is thus defined as

where F() is the galaxy spectral energy distribution (SED), and S() is the filter response function [\protect\astronciteKing and Ellis1985]. Frei and Gunn frei94 computed tables of galaxy colors (including B) and B k-corrections for galaxy types E, Sbc, Scd and Im. The (B) color and the B k-correction k in equation were computed by interpolating the Frei and Gunn frei94 tables to obtain two bivariate functions f and h such that B = f(,z) and k = h(,z). The functions f and h were both calculated with the IRAF SURFIT and two third-order interpolation schemes (Legendre and polynomial) as a consistency check. The observed Gunn color was used as an indicator of galaxy type. M values obtained with the Frei and Gunn frei94 tables were double-checked against those obtained in a much indirect way from the filter transformations and k-corrections given in King and Ellis kell85, Windhorst et al. wind91 and Metcalfe et al. met91. The systematic discrepancy between the two methods did not exceed 0.1 mag although the difference in M could reach 0.3 mag for some galaxies. The M values based on the Frei and Gunn frei94 tables were adopted as the true values because they were computed in a more direct way using more recent work.