COMMISSION 27 OF THE I. A. U.
           INFORMATION BULLETIN ON VARIABLE STARS
                        NUMBER 847

                                Konkoly Observatory
                                Budapest
                                1973 November 23


       A RECENT VERY LARGE PERIOD INCREASE IN U CEPHEI ?


   Recently Bakos and Tremko (1973) made the interesting observation 
that in 20 days in August 1969 the period of U Cephei increased 
by about 0.0015d. Such an increase is extremely large, as
they pointed out, corresponding to Delta P/P = + 6 X 10^-4. Additional
times of minimum, obtained since 1969, suggest that this very large
increase is not well established.

   The figure below is an O-C diagram based on the ephemeris
	JD(hel.) = 2440086.4915d + 2.493047d E + 7.8d X 10^-8  E^2.
The filled circles are the four times of Bakos and Tremko. The open
circles are two heretofore unpublished times of my own. The circled
points are the 16 photoelectric times later than JD 2440000 listed
in Table 14 of Batten (1973).

                             [FIGURE 1]

    My two new times of primary minimum are
      JD (hel.) = 2440644.9351d +/- 0.0002d
  and JD (hel.) = 2440679.8413d +/- 0.0003d
In both cases the error is based on the standard deviation of separate 
determinations in U,B, and V. But the external errors could
be larger, as much as approx. +/- 0.001d. The second time was determined 
by the tracing paper method. The first time was determined with the
help of the second, since only the falling branch was observed, and
hence should be relatively less reliable. To avoid an asymmetry
at the bottom of the light curve (rising branch brighter) the
lowest approx. 0.3m was not used.
   The problem is that the later limes continue to be satisfied
by the pre-1969 ephemeris. The broken line, which is the O-C curve
which would be required by the interpretation of Bakos and Tremko,
implies that three large period changes occurred within approximately 
100 days. The first change is drawn to occur, as they suggest, 
20 days before their last time of minimum. Such an interpretation 
is possible but would have the following implications which
should be appreciated:

   l. The three period changes are very large: Delta P/P = 7.5 x 10^-4,
-9.0 x 10^-4, +1.5 x 10^-4. These are between one and two orders of
magnitude larger than the alternate period changes of Delta P/P ~ 10^-5
which are characteristic of similar binary systems.
   2. The time scale of ~100 days is much shorter than the 5 - 10
year time scale for alternate period changes observed in U Cephei
in the past.
   3. After a sequence of three very large changes the period has
returned, seemingly by coincidence, to very nearly its original
value.
   The case for a large period increase rests on only one time of
minimum. Therefore (although admittedly the other three points of
Bakos and Tremko fit very well in the O-C diagram, their average
deviation being only a bit more than +/- 0.001d) a reasonable 
alternate interpretation would be that the last time is, for some reason,
in error by ~0.01d.


13 November 1973


                             DOUGLAS S. HALL

                             Astronomisches Institut
                             Universitat Tubingen
                             D-7400 Tubingen
                             Germany

References:
Bakos, G.A. and Tremko, J. 1973, Bull.Astr.Soc.
    Czechoslovakia 24, 298. [BIBCODE 1973BAICz..24..298B ]
Batten, A.H. 1973, Pub.Dom.Astrophys.Obs. 14, No. 10. [BIBCODE 1974PDAO...14..191B ]