4. General results

Due to the large size of the data set collected in this work, it is nearly impossible to show all the spectra in a traditional form. Instead, we will present our monitoring in a condensed way showing most of the relevant features that can be found in it. As examples, some individual spectra are displayed Fig. 1. The plots of all the data for the 21 objects of the monitoring can be found in Figs. 2 to 22; they are discussed in detail object by object in Sect. 5. In these plots we show our SiO data and other simultaneous observations at different wavelengths that are of interest for our discussion. These complementary data are optical curves for most of the objects in our sample, from the AAVSO International Data Base (Mattei, private communication), and infrared light curves (in L' [3.79 $\mu$m] and M [4.64 $\mu$m] filters), from Le Bertre (1993) for four objects: VY CMa, R Aqr, IK Tau, and IRC+10011. The SiO data are presented in two different forms. On the right side of each figure, we show contour plots of the SiO emission (flux density in Jy) as a function of both time (Julian date, JD) and velocity (LSR) for the two lines (v=1 on the bottom and v=2 on the top). These plots allow to follow the evolution of the individual SiO maser peaks and the changes of the line shape during the observing period. In these plots we also show the variation of the velocity centroid of the total emission (thick line). Between the v=1 and v=2 contour plots, two rows of small ticks mark the dates in which the observations were performed. Note that the contours in these figures are not linearly spaced and that the noise level, usually given by the first contour, is in most cases very small compared to the peak intensity. On the left side of each figure, in addition, we have made plots of the peak flux (Jy) and (velocity) integrated flux (Jy $\hbox{{\rm km$\,$ s$^{\rm -1}$ }}$) of the two studied SiO masers vs. time (Julian date, JD). These are the SiO light curves, to be compared with the optical and/or NIR light curves when available (displayed on the top-left corner). If possible, to ease this comparison, the epochs of optical maximum emission are shown in both SiO spectra and light curve plots by vertical lines. For the cases of IK Tau and IRC+10011, since no optical curve is shown, we have indicated the IR maxima. For OH 26.5+0.6 and Orion IRc2 no optical neither NIR maximum epochs are indicated.

The data corresponding to regular variables confirm that, in general, there is a correlation between the light curve of the observed SiO transitions and the optical variability, with a phase lag $\phi\sim~0.1-0.2$ of the first respect to the second (see, for example, Figs. 11, 12, and 18); see Martínez et al. (1988) and Hjalmarson & Olofsson (1979), for previous reports on this result. It is known that there is a similar phase lag between optical and infrared light-curves for the type of regular variables considered here (e.g. Lockwood & Wing 1971), so the SiO maser emission and IR flux vary in phase for these objects. This is clearly confirmed for R Aqr and IK Tau, for which NIR monitorings simultaneous to our observations are presented in Figs. 7 and 19. In the infrared source IRC+10011, that is very weak in the visible, the NIR flux has a regular variability (showing a period of $\sim$ 700 days) and the SiO maser emission also vary in phase with it (see Fig. 20). The red supergiant VX Sgr and the semiregular GY Aql could show the same type of behavior (Figs. 4 and 5). See next section for more details.

Even when a periodic behavior has been found in the observed SiO maser emission, the repetitiveness from one cycle to the next is poor, for instance the maximum and minimum intensities may significantly vary from one cycle to another. Sometimes, some expected SiO maxima do not appear or are very weak (see Figs. 14 and 17).

The line profile of the SiO masers has normally several components 1-2 km$\,$s$^{\rm -1}$ wide over a range of $\sim$ 10 km$\,$s$^{\rm -1}$ (giants) to $\sim~20-40$ km$\,$s$^{\rm -1}$ (supergiants). The line shape is generally similar in the two (v=1 and v=2) maser transitions. Also, we have detected sudden changes in the line shape of both lines in certain epochs (see e.g. Figs. 7, 12, and 14). No periodic change of the profile structure or velocity centroid has been found.