Yura Krugly, at Kharkov, Ukraine and Petr Pravec at Ondrejov, Czech Republic, independently determined a synodic rotation period of 2.7 hours from observations taken in May and June 1997, respectively.

During an (ongoing) observing campaign at ESO, La Silla, Chile, we (SM and GH) have been performing an extensive photometric monitoring of this object.
After the first few nights of observations it became clear that the object displayed a peculiar lightcurve behavior. In particular, while for most of the time the asteroid showed a typical two-maxima-two-minima pattern (like this ), sometimes it looked like this or this .

For sure this object deserved a closer look! Since Dionysus at this time is observable for only 4 hours at most from a particular site, we asked Petr Pravec at Ondrejov Observatory, whether he could resume his Dionysus observing program, which would allow us to achieve a continuous lightcurve coverage of nearly 7 hours. Petr agreed enthusiastically and a series of joint ESO/Ondrejov observations was started.

The hypothesis was made that the weird lightcurve behavior could be due to occultations/eclipses by a satellite orbiting the asteroid, with mutual occultations occuring roughly every 1.15 days. This hypothesis was already suggested to explain the irregular lightcurves of 1991 VH (Pravec et al. 1997, IAUC 6607) and 1994 AW1 (Pravec et Hahn 1997, Icarus 127, 431). This assumed period led to a prediction for the next occurence to be observed around June 7.0 UT.
Observations by Lenka Sarounova at Ondrejov confirmed this prediction by detecting an event on June 6.918 UT, while a further confirmation was obtained at ESO a day after, when another event was recorded on June 8.071 UT.
The analysis of all available light curve data from ESO and Ondrejov led, after substracting the unaffected "normal" lightcurve, to residual plots which closely resemble lightcurves from eclipsing binary stars. In the figures filled circles mark the calculated epochs when occultation event should have occured, based on a period of 1.155 days, which has been measured from the time difference between the observed lightcurve attenuation events.

With an occultation period of 1.15 days, one particular site can observe on average only one event every week. Due to sampling gaps in our data, there is still the possibility that the period is half of the mentioned value. The open circles in the residuals plot correspond to the predicted occultation dates in the case that the half period is correct.

In order to determine unambiguously the orbital period, and the parameters needed to model the assumed binary system, further observations are needed over a longer period and from different longitudes. During the forthcoming close approach the viewing geometry will change and the occultations/eclipses might eventually cease. Observations aimed at monitoring the development of the lightcurve behavior might help to further constrain the geometry of the asteroid/satellite system.