The merits of adding a polarimetric equipment to an optical interferometer has been
recognized as early as 1970's from the observation of a few hot stars by the Narrabri
intensity interferometer (Hanbury Brown 1974). Compared to polarimetric techniques, a
polarimetric interferometer can be thought of as a slide-caliper which measures
the distorsion of a scattering atmosphere seen through a linear polarizer
(Cassinelli & Hoffman 1975). In practice, the detection of this distorsion needs only
two series of observations: the measure of the fringe visibility through a
polarizer parallel to the interferometric baseline and perpendicular to it.
As pointed by Cassinelli and Hoffman this possibility works even when the polarization
of the atmosphere is axially symmetric -where classical polarimetry fails to detect
any effect. This applies a fortiori to the intrinsec oblate circumstellar envelopes
of Be stars, PMS objects or contact binaries (Poeckert & Marlborough 1978;
Fischer
et al. 1996; Vakili 1981).
Polarization measurements are also relevant for calibrating
systematic effects due to the differential rotation between incoming electromagnetic
fields from separated telescopes (Beckers 1990; Tinbergen 1988). The present work
aimed at comparing theoretical predictions of polarization effects in a stellar
interferometer to actual data on the sky (Rousselet-Perraut et al. 1996). It was
also initiated as part of an effort from our group to carry high spatial observations
in polarized light with the GI2T new beam-combiner REGAIN (Mourard et al. 1994a). The
bright Be star 
Cas presents strong polarization properties a priori
(Clarke
1990) and corresponds to a suitable target for evaluating the limits of such
observations on the GI2T. In the following section we describe the instrumentation
and data analysis which have been used for our study. The next section describes the
interpretation of our interferometric measurements based on a simple model for the
extended envelope of 
Cas. The angular diameter of the envelope in the
continuum is also compared to previous results of 
Cas by the Mark III and
the GI2T in the H
emission line (Quirrenbach et al. 1993; Stee et al. 1995).
In the last section our analysis is used to constrain the physics of the envelope
of this star. Finally we conclude on a few scientific applications of the future
polarimetric mode of GI2T + REGAIN.