COMMISSION 27 OF THE I. A. U.
           INFORMATION BULLETIN ON VARIABLE STARS
                        Number 3486

                                         Konkoly Observatory
                                         Budapest
                                         28 June 1990

                                         HU ISSN 0374 - 0676

             ATMOSPHERIC ECLIPSE OF delta SGE

   The long period (P=3720 days) bright spectroscopic binary
delta Sge exhibits complex spectroscopic phenomena shortly before and
after the conjunction times (McLaughlin et al. 1952; Batten and
Fisher, 1981). Such phenomena were thought to be due to an
atmospheric eclipse in the system during the conjunction times.
Based on the timing of such phenomena and the period of the system
the last conjunction must have occurred around the beginning of
May 1990. However, the improved parameters of the system by
Reimers and Schroder (1983) suggest no eclipse during the
conjunction times. Because the longitude of periastron is very close
to 270deg, conjunctions are expected to be very near to periastron
passage. The last periastron passage was expected to be in the
second week of April 1990.
   It is possible that the B9 V component may be eclipsed by the
extended atmosphere of the M2 II component at conjunction. If so,
then any eclipse (total or partial) in the system could be detected
photometrically in short wavelengths, because the contribution of the
early-type component to the total light from the system is only a few
percent, but it increases towards shorter wavelengths.
   We have observed the system with UBV filters on 21 nights in
1988, and 85 nights in 1989. The differential observations with
respect to the comparison star beta Sge were made by using an
EMI 9789 QB photomultiplier attached to the 30 cm Maksutov
telescope of Ankara University Observatory. The differential
brightness measurements of beta Sge with respect to the check star
alpha Sge were found to be sensibly constant during the observations :
DeltaV= -0.00+/-0.01, DeltaB=-0.27+/-0.02, and DeltaU=-0.73+/-0.04. The individual
differential magnitude determinations were corrected for differential
atmospheric extinction, and light time effect. The 1988 observations
were already reported (Derman, et al., 1989). The 1989 observations
in the sense variable minus comparison are plotted in Figure 1
against the heliocentric Julian date.


                            [FIGURE 1]

 Figure 1. The differential U, B and V observations of delta Sge
 against heliocentric Julian date (HJD +2400000) in 1989


   New observations in Figure 1 verify the periodic nature of the
light variation. The amplitude of the variation seems to decrease first
until May 1989 and then increases until the end of July 1989. while
the period decreases steadily from about 60 days to 28 days. Three
cycles of the light variation are clearly seen in V and B observations in
Figure 1. The U observations show almost no sign of periodic
variation. Thus, the periodic variations should be related to the late
type giant component, because this component has large contribution
in V and B but less in U magnitudes. Three cycles observed between
May and August 1989 have decreasing periods of about 37, 30 and 28
days, respectively. Total mean light from the system increases
steadily from May to August 1989, then drops sharply to a lower
level, indicating the beginning of an atmospheric eclipse. Contrary to
the expectation such behavior of the light variation is better observed
in longer wavelengths (see V and B light curves in Figure 1). If the
loss of light starting from August 1989 is due to an atmospheric
eclipse then the eclipse duration may be as large as 18 months.
because the eclipse begins about nine months before the estimated
conjunction time. In this case the inclination of the orbit or the size
of the giant component ought to be much larger than the predicted
values. Otherwise the period of the system may be shorter than 3720
days as suggested by Batten (1989). Figure 1 suggests that the hotter
component is eclipsed from the beginning of August 1989 by the
extended atmosphere of the cool giant component. The periodic
oscillations of the total light fade but do not disappear completely
during the eclipse phases. Further observations may reveal the
duration of eclipse phenomena which will bring important
constraints on the system parameters.


                       OSMAN DEMIRCAN
                       ETHEM DERMAN
                       AYVUR AKALIN

                       Ankara University Observatory
                       Science Faculty, Tandogan
                       06100 Ankara / Turkey


References

Batten, A.H., Fisher, W.A.: 1981, Publ. Astron. Soc. Pacific. 93. 769. [BIBCODE 1981PASP...93..769B ]
Batten, A.H.: 1989. I.B.V.S. No. 3357.
Derman, E., Demircan, O., Engin, S., Akalin, A.: 1989, 
    I.B.V.S. No. 3343.
McLaughlin, D.B., Weston, E.B., Chadwick, M.: 1952, Publ. Astron.
    Soc. Pacific. 64, 300. [BIBCODE 1952PASP...64..300M ]
Reimers, D., Schroder, K.-P.: 1983, Astron. and Astrophys., 124,
    241. [BIBCODE 1983A&A...124..241R ]