Bayesian statistical methods of Gregory-Loredo and the Bretthorst generalization of the Lomb-Scargle periodogram have been applied for studying activity cycles of the early K-type subgiant star CF Oct. We have used a ~45 year long dataset derived from archival photographic observations, published photoelectric photometry, Hipparcos data series and All Sky Automated Survey archive. We have confirmed the already known rotational period for the star of 20.16 d and have shown evidences that it has exhibited changes from 19.90 d to 20.45 d. This is an indication for stellar surface differential rotation.The Bayesian magnitude and time--residual analysis reveals clearly at least one long-term cycle. The cycle lenght's posterior distributions appear to be multimodal with a pronounced peak at a period of 7.1 y with FWHM of 54 d for time-residuals and at a period of 9.8 y with FWHM of 184 d for magitude data. These results are consistent with the previously postulated cycle of 9+/-3 years.

The successful detection is reported of radial-velocity variations due to orbital motion of the substellar companion of the star tau Bootis, from data obtained with a small aperture (0.4m) telescope and a fibre-fed high-resolution spectrograph. Radial-velocity observations from observing runs in 2000 and 2004 reveal a periodic variation of 3.30 +/-0.02d, which is consistent with the previously determined value of 3.3125 +/-0.0002d. We fit our data to a circular orbit with the known period, and derive a velocity amplitude of 471 +/-10m s-1 (in agreement with the previously published value of 469 +/-5m s-1), and determine a time of maximum velocity (Tmax) of HJD 2453113.95 +/-0.01. These observations explore the minimum system requirements for precise radial-velocity measurements.

Saturated stellar images on digitised photographic plates are many times greater in area than the `seeing disk' seen in unsaturated CCD images. Indeed the flux profile of a bright star can be traced out for several degrees from the star's centre. The radius of the saturated stellar image can often be directly related to the magnitude of the star, a fact well known and exploited in iris photometry. In this work we compare the radial flux profile of stars in the approximate range B ~9 to ~13 mag, obtained from scans of plates from the Bamberg Sky Patrol archive, with a profile of the form measured by King. We show that simple aperture photometry of saturated stellar images, obtained from photo-positives of scanned photographic plates, yield data that are in agreement with simulations using a (saturated) synthetic stellar radius profile. Raw plate magnitudes from this aperture photometry can be easily and satisfactorily transformed to standard magnitudes, as demonstrated in a recent study carried out by the current authors.

We use the archive of the Bamberg Sky Patrol to obtain light curves of the active K subgiant CF Octantis for the interval 1964-1976. Digitised images of the field near CF Oct were obtained with a flat-bed scanner. Aperture photometry was performed of photo-positives of these images. Using a transformation to second order in plate magnitude, and first order in B-V, for 9 field stars for each plate, the B magnitudes of CF Octantis were obtained for just over 350 plates. The estimated precision of an individual determination of the B magnitude of CF Oct is 0.05 mag. Analysis of the resulting data reveals the known 20 d rotational variation of this star, and shows the evolution of the light curves from year to year. We obtain light curves with good phase coverage for 1964 to 1969 inclusive, partial light curves for 1970 and 1976, and a few data points from 1971. The amplitude of variation ranges from ~0.2 to \~0.4 mag. There is evidence that the characteristic rotation period of the star in the 1960s was slightly less than that measured from photoelectric photometry in the 1980s.