Spatial information is crucial in studies of the internal kinematics of intermediate redshift galaxies. The line emission must be spatially-resolved to ascertain that the [OII] kinematics is actually coupled to the rotation (i.e. mass) of galaxies. It may not always be the case. For example, the nucleus of a galaxy could be undergoing massive star formation, and the [OII] emission could all be concentrated in the nucleus. Winds from a few giant HII regions could dominate the [OII] line widths making them unusable as virial measurements of galaxy mass. Rotation can be readily seen in spectra with good spatial information as a ``S-shaped'' or ``tilted'' line emission. The use of the Tully-Fisher relation as a comparison will only be valid for galaxies in which a rotational component has been identified and isolated from other types of motion (random, radial, etc.) which may be present.
Low-order, ``tip/tilt'' image stabilization systems have been in operation at the Canada-France-Hawaii Telescope (CFHT) for a few years now. The Subarcsecond Imaging Spectrograph (SIS) is the latest example of such a system (see section 
). These systems improve the seeing FWHM by 0-0and routinely deliver images with 0FWHM seeing. It was thus possible to obtain spatially resolved spectra of relatively compact (effective radius 
1 
-3pt ) intermediate redshift galaxies. CFHT was ideally suited for a survey of the internal kinematics of field galaxies with redshifts in the range 0.25
0.45 using the [O II] 
37263729 Åemission line doublet as a kinematical tracer. The availability of the CNOC cluster survey database (see section ) which contained information on a large number of galaxy candidates suitable for internal kinematics observations greatly facilitated target selection.
The survey was aimed at answering a number of questions: (1) What kind of kinematics do intermediate redshift galaxies have?, (2) What are the masses of intermediate redshift galaxies?, (3) Is there a systematic shift in the intermediate redshift Tully-Fisher relation as described in section ?, (4) Is the TF slope the same as the local one or equivalently, is the shift from the local TF mass-dependent? and (5) Do galaxies with the same mass undergo different amounts of luminosity boosting? If yes, what is (are) the other underlying parameter(s)?
The answer to question 1 may not be as obvious as one might think at first sight. Very little was known about [OII] kinematics in intermediate redshift galaxies at the time this survey was started, and yet Question (1) lay at the heart of any proposal to use internal kinematics to study galaxy evolution. A wide range of [OII] kinematics would indicate that a number of processes are responsible for the star formation activity seen at intermediate redshifts. Questions (3) and (4) address what kind of evolution excess galaxies might be undergoing. Luminosity-dependent luminosity evolution predicts that some L
galaxies should be low-mass galaxies significantly shifted from the local TF relation whereas other L galaxies should be massive galaxies with little or no deviation from the local TF relation. A systematic TF shift preserving the slope of the local relation would rule out luminosity-dependent luminosity evolution. An answer to Question (5) may elude the present survey as a larger sample will probably be required to explore the dependence of luminosity boosting on various galaxy properties.