We study three representative, overactive spotted K giants (IL Hya, XX Tri, and DM UMa) known to exhibit V-band light variations between 0.65-1.05 mags. Our aim is to find the origin of their large brightness variation. We employ long-term phase-resolved multicolor photometry, mostly from automatic telescopes, covering 42 yr for IL Hya, 28 yr for XX Tri, and 34 yr for DM UMa. For one target, IL Hya, we present a new Doppler image from NSO data taken in late 1996. Effective temperatures for our

more »

... argets are determined from all well-sampled observing epochs and are based on a V-I_C color-index calibration. The effective temperature change between the extrema of the rotational modulation for IL Hya and XX Tri is in the range 50-200 K. The bolometric flux during maximum of the rotational modulation, i.e., the least spotted states, varied by up to 39% in IL Hya and up to 54% in XX Tri over the course of our observations. We emphasize that for IL Hya this is just about half of the total luminosity variation that can be explained by the photospheric temperature (spots/faculae) changes, while for XX Tri it is even about one third. The long-term, 0.6 mag V-band variation of DM UMa is more difficult to explain because little or no B-V color index change is observed on the same timescale. Placing the three stars with their light and color variations into H-R diagrams, we find that their overall luminosities are generally too low compared to predictions from current evolutionary tracks. A change in the stellar radius due to strong and variable magnetic fields during activity cycles likely plays a role in explaining the anomalous brightness and luminosity of our three targets. At least for IL Hya, a radius change of about 9% is suggested from m_bol and T_eff, and is supported by independent vsin(i) measurements.