COMMISSION 27 OF THE I. A. U.
         INFORMATION BULLETIN ON VARIABLE STARS
                      Number 3383

                                        Konkoly Observatory
                                        Budapest
                                        9 October 1989

                                        HU ISSN 0374 - 0676


    ON THE INSIGNIFICANCE OF SOME CLAIMED OSCILLATION PERIODS


The recent IBVS entitled "Detection of new short period oscillations in
PG1711+336 (V795 Her)" by Ashoka et al. (IBVS No. 3352) demonstrates an
elementary misunderstanding of the significance or otherwise of apparent
periods in the light curves of cataclysmic variables (CVs). The short section
of light curve of V795 Her obtained and illustrated by the authors shows the
rapid brightness variations typical of CVs. Without implying any physical
connection with solar or stellar flares, let us call the brightenings and
fadings in the light curve "flares". The flares occur on a variety of
timescales. If at least some fraction of the flares can be shown to occur
repetitively, for many cycles, with a well defined period, then the claim of a
periodicity can be made, and it is worthwhile enquiring the physical reason for
such an eigenmode in the system. An extension to quasi-periodic oscillations
can be made if there are sufficient oscillations that can be shown to occupy a
relatively narrow range of period in the Fourier transform.

However, to claim, as Ashoka et al. do, that the occurrence of just three
flares spaced nearly equally in time (during the time 20.23 to 21.33 of their
Figure 1) constitutes the detection of a periodicity of 1389 secs, or even
(because it was not present in the second half of their observing run) that it
is quasi-periodic, is nonsense.

In a paper in press (A.W. Shafter, E.L. Robinson, D. Crampton, B. Warner
and R.M. Preston - Ap.J. submitted) a total of 22 hrs of high speed
photometry of V795 Her is presented: power spectra show no persistent short
period oscillations. However, if the light curve were to be subdivided into
sections as short as those used by Ashoka et al., any number of apparent
"oscillation periods" would be found. This is true of most CVs. For example,
the reader is invited to inspect the light curve of U Geminorum just prior to
eclipse (p. 96 of I.A.U. Symposium No. 73): superficially there appears to
be a periodicity in the flares, but power spectral analysis shows that there is
none. If analysed in short sections, however, many apparent short-lived
periodicities would emerge. There are hundreds of metres of plots of other CV
light curves that could be similarly overinterpreted.


BRIAN WARNER
Department of Astronomy
University of Cape Town
7700 Rondebosch
South Africa