COMMISSION 27 OF THE I. A. U.
             INFORMATION BULLETIN ON VARIABLE STARS
                         Number 3569

                                                 Konkoly Observatory
                                                 Budapest
                                                 7 March 1991

                                                 HU ISSN 0374 - 0676


       RECENT MINIMA AND IMPROVED EPHEMERIDES FOR THE
        ECLIPSING HOT SUBDWARFS LB3459 (=AA DOR) AND
                    BD-7d3477 (=HW VIR)

   The hot sdOB star LB 3459 (=AA Dor) was discovered to be an
eclipsing binary by Kilkenny et al. (1978) and has a rather short
period of 6h17m. The parameters of the secondary star are
not well known but it appears to be close to a degenerate 
configuration (Kudritzki et el 1982) and most of the visible light
from the secondary is apparently reflected primary light. The
sdB star BD-7d3477 (=HW Vir) was discovered to be an eclipsing
binary by Menzies & Marang (1986) during a survey of UV-bright
objects. It has a very short period (2h48m) and the
secondary could be a red main-sequence star (Menzies & Marang
1986).
   Because of the evolved nature of the primary stars in these
systems end the short periods, it seems highly likely that both
could have passed through at least one "common envelope" stage
(Paczynski 1980). Currently, both binaries appear to be
detached so that mass exchange by Roche lobe overflow is not
likely to be occurring. Angular momentum loss by gravitational
radiation or by mass loss vie stellar winds would effect the
periods; the former should only be measurable after ~10^2-10^3
years (see, e.g. Paczynski 1967) but the letter might be more
effective. We have therefore made occasional observations of
eclipses of both systems with a view to establishing very
accurate ephemerides to test for period changes.
   The most recent ephemerides for AA Dor and HW Vir have been
given by Kilkenny (1986) end Marang & Kilkenny (1989) respectively. 
The new timings of primary minima together with
estimated errors are listed in Table 1, where the cycle numbers
,n, are determined from the above mentioned ephemerides. All
observations were made with the 1.0m and 0.5m telescopes at the
Sutherland site of the South African Astronomical Observatory.
Various filters were used (see Table 1) with 20- or 30-second
continuous integrations to obtain the eclipse curves. Sky
measures were made before and after eclipse and usually near mid-
eclipse; the sky corrections were then interpolated for all
eclipse date.
   The results for HW Vir are more accurate than those for AA
Dor, probably due to the fact that HW Vir is rather brighter than
AA Dor and that the latter is a foreground LMC object which means
a smaller aperture is necessary and so scintillation and guiding
effects are more serious.
   Combining the new eclipse timings with published results, we
obtain for HW Vir:

   T_min = 244 5730.556071 + 0.1167196336n
                +-0.000014 +-0.0000000017

from 34 eclipses. T_min is the time of primary mid-eclipse and
the errors are formal errors from a linear least-squares fit to
the times of mid-eclipse.

   For AA Dor we obtain

   T_min = 244 3196.348685 + 0.2615397198n
                +-0.000016 +-0.0000000017

from 27 eclipses. The eight timings between cycles 1000 and
4000 are excluded from the linear least squares solution because
these were either multicolour (uvby) and therefore of much poorer
time resolution or were of poor quality. Including these data
in the solution makes almost no difference to the ephemeris
coefficients but does make the errors significantly worse.


             TABLE 1 New timings of primary minima

                    HJD            Est.error     n    Filter    Tel

AA Dor         2446795.39692   +-  0.00005     13761    V       1.0m
(=LB3459)         6801.41230       0.00005     13784    V       1.0
                  7121.53695       0.00010     15008    y       1.0
                  7850.44822       0.00005     17795    V       1.0
                  8267.34250       0.00003     19389    V       1.0

HW Vir         2447968.53836   +-  0.00002     19174    B       0.5m
(=BD-7d3477)      7972.50682       0.00002     19208    B       1.0
                  8267.57410       0.00002     21736    V       1.0


                      [FIGURE 1]

 Fig. 1. (O-C) diagrams for AA Dor and HW Vir basd on the
        ephemerides determined in this paper. The AA Dor data
        were obtained between 1977 and 1991; the HW Vir data
        between 1984 and 1991.


   Figure 1 shows the (O-C) residuals for both binaries. In
the case of AA Dor, the linear fit is very good and there is no
evidence for period change. For HW Vir there is a weak indication 
in the most recent data that the linear fit may not be good
enough.
   It is unfortunate that no observations of HW Vir were made
during 1986-1989 and it is clear that further data are very
desirable.


                D. KILKENNY, M. HARROP-ALLIN, F. MARANG

                South African Astronomical Observatory
                PO Box 9, Observatory 7935, S. Africa



References:

Kilkenny, D., Hilditch, R.W. & Penfold, J.E., 1978. Mon. Not.
     R. astr. Soc., 183, 523. [BIBCODE 1978MNRAS.183..523K ]
Kilkenny, D., 1986. Observatory 106, 160. [BIBCODE 1986Obs...106..160K ]
Kudritzki, R.P., Simon, K.P., Lynas-Gray, A.E., Kilkenny, D. &
     Hill, P.W., 1982.  Astr. Astrophys., 106, 254. [BIBCODE 1982A&A...106..254K ]
Marang, F. & Kilkenny, D., 1989. Inf. Bull. Var. Star, 3390.
Menzies, J.W. & Marang, F., 1986. IAU Symposium 118, 305. [BIBCODE 1986IAUS..118..305M ]
Paczynski, B., 1967. Acta. Astr., 17, 287. [BIBCODE 1967AcA....17..287P ]
Paczynski, B., 1980. Acta. Astr., 30, 113. [BIBCODE 1980AcA....30..113P ]