2 The observations

 

Figure 1: The $\vec E$-vectors of the UBV polarization measurements. The length of each line is proportional to P, the percentage polarization and the angle at which each line is drawn relative to the galactic north pole in the direction of increasing longitude is $\theta_{\rm gal}$

The observations were carried out at the Danish 1.5 m telescope at La Silla, on 10 nights in 1990 Jan. 08 - 17, 7 nights in 1991 Jan. 18 - 24, and 3 nights in 1992 Apr. 22 - 24. A multichannel version of the double image chopping polarimeter (Korhonen et al.1984) was used, and the measurements in different colours were collected simultaneously by using dichroic filters to split the light into the spectral regions. The resulting passbands are close to the UBV system, with effective wavelengths 0.36, 0.44, and 0.54 $\mu$m, respectively. The sky background polarization was directly eliminated by using a plane parallel calcite plate as the polarizing beam splitter, and the intensities of the two beams are measured, for each colour, by a single photo-multiplier using chopping techniques (Piirola 1973). As determined from observations of unpolarized standard stars (Tinbergen 1979), the instrumental polarization was found to be very small. The zero point of position angles was determined by observations of seven standard stars with large interstellar polarization. These standard stars were selected from the list suggested by Serkowski (1974) and they were measured regularly throughout the observing periods. For most of the cases, the observation of one star took about 15 minutes and consisted of four sets of measurements at eight position angles of the instrument. At each position angle, integrations of 10 seconds were made for each one of the two light beams. In order to improve our accuracy, six sets of measurements were collected for some of the faintest stars.

The targets were selected from the Knude's A- and F-type star catalogue (Knude 1977, 1978). The observed stars belong to 35 low galactic latitude ($\vert b\vert \le 30\hbox{$^\circ$}$) Kapteyn's Selected Areas and cover the third and fourth quadrants of the galactic plane.

 

Figure 2: Estimated error in the B-band versus stellar visual magnitude. The errors of bright stars ( ) are mainly dominated by atmospheric scintillation. For fainter stars, it is dominated by photon shot noise and, in the present case, may be roughly represented by $\sigma_{ B} = 5.5 \ 10^{-4}(10^{ V})^{1/5}$ (solid curve)