COMMISSION 27 OF THE I.A.U.
               INFORMATION BULLETIN ON VARIABLE STARS
                            Number 1500

                                                      Konkoly Observatory
                                                      Budapest
                                                      1978 November 8


A SPECTRAL ANALYSIS OF THE LIGHT CURVE OF THE RV Tau STAR EP Lyr

As it is well known, the maximum entropy method is so far the
powerful one to estimate the power spectra of time series with equally 
spaced data (Percy, 1977 and references therein).
   This method has been applied to analyse the light curve of
the RV Tau star EP Lyr, since of this star there is a long run
of photoelectric observations, which has sufficient continuity to
be uniformly sampled (Wachmann, 1968).
   The power spectrum has been computed using the Burg recursive
procedure and adopting the Akaike criterion to choose the length
of the error prediction filter (Richer and Ulrych, 1974).
   In Fig. 1 we have the spectrum, which has in abscissae the
frequencies (cycles to day), and in ordinates the logarithm of the
power. This spectrum has been computed from the V observations
made between J.D. 2439614 and J.D. 2939814. The data have been
connected with a free hand curve and then this has been sampled
with a step of 5 days.
   It can be seen that the formal period of the star yields a
series of peaks, which are due to its harmonics (the order of each
one is written above the downward arrows). The period, which has
been determined fitting these harmonics with the least squares,
is of 84.2 +- 2.5 days, which is in good agreement with the one
of 83.315 given by Wachmann, who determined it using a very long
series of minimum epochs. Moreover there are two other peaks 
(upward arrows) at the frequencies of about 0.023 and 0.067 c/d. These
peaks are the responsible of the not perfect regularity of the
light curve. The higher of these two peaks is very close to the
one due to the 2 deg harmonic of the formal period; so it is bad
resolved. Its position can be estimated only with approximation,
since in such cases the peaks tend to depart one from the other.

                      [FIGURE 1]
                      [FIGURE 2]

   A similar consideration can be made for the second peak. So
probably this one is third harmonic of peak at 0.023 c/d. In
such hypothesis it results a second periodicity of 45.1 +- 1.7
days. Unfortunately there are not in the literature other 
sufficiently long runs of observations to confirm this second 
periodicity, which till now passed unnoticed owing to its proximity 
to the 2 deg harmonic of the formal period.
   Wachmann gives also a series of epochs of principal minima
(which cover 160 formal periods) and the concerning (O-C), which
have been computed with the period P=83d.315. The spectrum of
this curve, which has been obtained with the same procedure
used for the V light curve, has been also computed (Fig. 2, in
abscissae the frequencies are in cycles to period).
   All the peaks in the spectrum are harmonics of a very
slow oscillation of 7067 +- 88 days. For this oscillation Wachmann 
gives a period of 3500 days, which as it can be clearly
seen from the spectrum, is only the 2 deg harmonic of the true
period.


                       LUCIANO MANTEGAZZA

                    Osservatorio Astronomico
                        Via Bianchi, 46
                       22055 Merate (Italy)



References:

Percy, J.R., 1977,  Mon. Not. R. Astron. Soc., 181, 647 [BIBCODE 1977MNRAS.181..647P ]
Richer, H.B., Ulrych, T.J., 1974, Astrophys. J., 192, 728 [BIBCODE 1974ApJ...192..719R ]
Wachmann, A.A., 1968, Abh. Hamburg. Sternw. Bergedorf, VIII, 4, 105 [BIBCODE 1968AAHam...8..105W ]