2 Observations

The OH observations were made with the Nançay transit radio telescope which has a half beamwidth of 3.5' in $\alpha$ by 18' in $\delta$ at 18 cm wavelength. The system temperature was about 50 K. The ratio of flux to antenna temperature was 1.1  ${\rm ~Jy~K^{-1}}$ at 0$^\circ$ declination. A frequency switching mode was used. Each observation lasted about one hour and achieved a mean rms of 0.03 Jy. The observations were made in both circular polarizations. Using a 1024 channel autocorrelation spectrometer, the first set of observations was obtained between February 1980 and December 1992 covering from 1 to 7 consecutive cycles depending on the star. During this period the autocorrelator was split into four banks of 256 velocity channels each, and the velocity resolution was about 0.14 km s^-1 in the main lines and 0.145 km s^-1 in the satellite line. Also, between February 1980 and March 1983 some linear polarization observations were also made for R Aql. For these observations the back end consisted of four banks of $64 \times 6$ kHz filters leading to a velocity resolution of about 1 km s^-1.

The second set of observations was obtained between January 1993 and January 1996, covering between one and two cycles depending on the star. For R LMi, the autocorrelator was split into four banks of 256 channels each leading to the same velocity resolution as previously mentioned. For the other sources, the autocorrelator was split into two banks of 512 channels each, thereby doubling the velocity resolution (i.e., $\simeq$ 0.07 km s^-1).

According to observations made on the calibration source W12, antenna temperature values are estimated to be correct to $\pm$ 18% between 1980 and 1984 and to $\pm$ 8% for later observations. Thus, between 1980 and 1984, degrees of polarization lower than 0.1 should be taken with caution.