2 Observations

All the observations reported in this paper were carried out with the Medicina 32-m radiotelescope during a number of sessions in the period January 1993-April 2000. A description of the telescope and the equipment is given by Comoretto et al. (1990), and here we will only give a summary of the main features.

At the frequency of the 6₁₆-5₂₃ transition of H₂O (22.23507985 GHz), the HPBW of the antenna is 1.9 arcmin. The pointing model was checked at the beginning of each 2-3 week session by maximizing the signal from a set of strong galactic H₂O masers (W3OH, Orion KL, Sgr B2, W49N). The resulting accuracy was always better than 25 arcsec. The antenna gain as a function of elevation was determined regularly by doing short ($\sim$10 min) total power integrations on DR21 (adopted flux density 18.8 Jy, Dent 1972). For each observing day, all gain measurements as a function of elevation were fitted with a polynomial curve, which was then used to calculate the conversion factor from antenna temperature to flux density for the spectra observed on that day. The calibration error resulting from the dispersion of the single measurements from the fit turns out to be 19%. On a few dates, no gain curve was measured: in this cases we have applied the closest gain curve in time and we estimate a corresponding calibration uncertainty of 7%. Therefore, a conservative estimate for such an uncertainty has to be taken equal to 21%.

Observations were always done in total power mode, with 5 min integration time on- as well as off-source. Depending on weather conditions, elevation, and spectral resolution, this resulted in 1$\sigma$ rms noise levels between 0.3 Jy and 6 Jy. The distribution of the rms noise for the 426 observations (including multiple observations of the same source) is shown in Fig. 1. The peak occurs at 0.7 Jy both for positions where emission was (solid line) and was not (broken line) detected. Compared to the observations presented in U1, there is a factor of $\sim$2 improvement in the rms of the present survey. In general, the band was centered on the expected velocity, either the velocity of the maser as reported in the literature, or, if available, the velocity of the molecular cloud in which the source is embedded.

Figure 1: The distribution of the 1$\sigma$ rms noise for the 426 spectra observed in U2. The bin size is 0.25 Jy