COMMISSION 27 OF THE I. A. U.
           INFORMATION BULLETIN ON VARIABLE STARS
                        Number 3413

                                           Konkoly Observatory
                                           Budapest
                                           3 January 1990

                                           HU ISSN 0374 - 0676


      VARIABLE POLARIZATION IN A COMPARISON STAR OF V3885 Sgr

   In 1983 polarimetric measurements of the novalike system V3885 Sgr
were performed at the European Southern Observatory (Metz, 1989). In order
to determine the interstellar polarization of the system (Haefner, Metz,
1982) the linear as well as the circular polarization of several comparison
stars neighbouring V3885 Sgr was determined several times. Whereas V3885
Sgr did not show any indication of intrinsic polarization (Metz, 1989), one
of the measured comparison stars proved to be variable in both, its linear
and circular polarization. The star (with the internal designation 11 on
the finding chart, used during the observation) is separated by roughly 10
arcminutes from V3885 Sgr the coordinates of the latter are:
alpha_2000= 19h47m40s, delta_2000=42d00'10" (2000). Star No. 11 was also 
measured by Bond (1978)
and is described in the literature as "Bond's star" with a brightness of
m = 10.6 and a spectral type of F2. Its position can be seen from the
finding chart of V3885 Sgr (Fig. 1). After eliminating all erroneous 
measurements - caused above all by a heavy snow storm at La Silla - altogether
43 linear and 33 circular polarization determinations of the comparison
star No. 11 remained for evaluation.
   The result of the determination of the linear polarization is:
Mean intensity in arbitrary units I = 60026, SDV (I) = 3972 (SDV is the
standard deviation for a single measurement)

Pmx = -0.0041,         SDV (Pmx) = 0.0010
Pmy = -0.0038,         SDV (Pmy) = 0.0012
Pm  =  0.0057          SDV (Pm ) = 0.0011
(Pmx, Pmy, Pm are the means of 43 single determinations of Px, Py, P)
   The result of the 33 circular polarization measurements is:

Mean intensity I = 57917, SDV (I) = 3621

Pmv = -0.0002, SDV (Pv) = 0.0018.


                             [FIGURE 1]

Figure 1. The finding chart used for V3885 Sgr and its comparison stars


  Since each single measurement was the result of 16 rotation steps of
the phase plates used in the old double beam ESO polarimeter and since the
integration lasted 10 sec for each of the steps the error of a single 
measurement caused only by photon statistics was of the order of 0.0007. This
is about a factor of 15 smaller than the derived standard deviations of the
measurements.

   The measurements of V3885 Sgr and its comparison stars lasted for 33
days and were interrupted two times by a break of more than a week. Therefore 
it was clear that the number of measurements of star No. 11 would not
be sufficient to carry out a precise periodogram analysis or to make further 
reductions. However one thing was striking immediately after a short
inspection of the data:

   Completely in contradiction to V3885 Sgr and all other comparison
stars the polarimetric errors of star No. 11 were much higher than derived
from the photon statistics. The high standard deviations in both the linear
and the circular polarizations indicated that star No. 11 was variable.
Though the quantity of data was rather small to make any further statements
concerning the comparison star, which had only occasionally been measured,
a periodogram was performed in order to find (with all precautions 
described above!) any periodicity in the observed variations of the 
polarization. To detect most of the pseudo periodicities necessarily caused by
numerous observational gaps, above all in observing the comparison stars,
the polarization values were replaced by random numbers with the same standard 
deviation as registered for the star No. 11 and analysed in the same
way. The instantaneous linear polarization degree P and the circular 
polarization degree PV were inspected in the range 1h - 120h with time intervals
of one hour. The reason for selecting this time interval for the periodogram 
was that a preliminary plot gave indications that, if at all, a periodic 
variation should lie in that interval. For the original as well as the
artificial data a Fourier analysis was made and gave the following result
for comparison star No. 11:

   In the linear polarization a period of 86 hours was present whereas
the circular polarization varied with a period of 87h. With respect to the
selected time resolution and considering further the small quantity of data
the results seem to be identical. Quite evidently this result has to be
checked by a sufficient number of further measurements.





                             KLAUS METZ

                             Universitats-Sternwarte
                             Scheinerstr. 1
                             D-8000 München 80
                             Germany

References:

Bond, H.E.: 1978, Publ. Astron. Soc. Pacific 90, 216. [BIBCODE 1978PASP...90..216B ]
Haefner, R., Metz, K.: 1982, Astron. Astrophys. 109, 171. [BIBCODE 1982A&A...109..171H ]
Metz, K.: 1989, IBVS No. 3385.