2 Observations and data analysis

 

2.1 Observational procedure

The observations were performed in February 1996 with the 20-m radio telescope of Onsala Space Observatory in Sweden. The half-power beamwidth of the telescope is 40$^{\prime\prime}$ at the CS and C³⁴S frequencies (97981.00 MHz and 96412.98 MHz, respectively) and 35$^{\prime\prime}$ at the CO frequency (115.3 GHz). The front-end was a 3 mm SIS receiver. The single side band system temperature referred to outside the atmosphere was 400 - 1000 K for CS, 200 - 300 K for C³⁴S and 900 - 1900 K for CO observations depending on the weather conditions and source elevation. The backend was a 256 channel filterbank with 250 kHz resolution.

The observations were made mainly in the frequency switching mode with the frequency offset of 15 MHz. The CS spectra at the peak positions and a part of the C³⁴S spectra were obtained using dual beam switching with a beam throw of 11$.\mkern-4mu^\prime$5. Initially, $3\times 3$ point CS maps with 40$^{\prime\prime}$ spacing were obtained for all sources. Then, the mapping was continued according to the initial results and extended normally to $\sim$10% of the peak antenna temperature value. Pointing was checked periodically by observations of nearby SiO masers; the pointing accuracy was $\sim 5$$^{\prime\prime}$.

The standard chopper-wheel technique was used for the calibration. We express the results in the units of main beam brightness temperature ($T_{\rm mb}$) assuming the main beam efficiency ($\eta_{\rm mb}$) of 0.56 and 0.50 at the CS and CO frequencies, respectively (L.E.B. Johansson, private communication). The temperature scale was checked by observations of Orion A, which we observed also at SEST (Paper I). The comparison with the SEST data shows that the main beam temperatures in these lines agree to within a few percent at the two instruments (we have taken into account the different spectral resolutions in our SEST and Onsala observations).

2.2 Source list

Our source list is based on the Arcetri Atlas of galactic H₂O masers (Comoretto et al. 1990; Palagi et al. 1993; Brand et al. 1994), which covers the declination range $\delta \gt -30\hbox{$^\circ$}$.The present observations were limited to the longitude range $l=120^{\circ} -212^{\circ}$ (approximately) and $\delta \gt 0$. We managed to observe almost all masers from this atlas satisfying these criteria; their coordinates are given in Table 1. A few objects (S 199, S 255) observed earlier at Metsähovi (Zinchenko et al. 1994) were omitted from the current list. NGC 281 was observed in C³⁴S only since it was not detected in this line at Metsähovi.

We designate all the sources according to their galactic coordinates. The common identifications with some well known objects are given in the last column.

  

Table 1: Source list

2.3 Association with the IRAS sources

  We searched the IRAS point source catalogue for possible associations with our targets. In Table 2 (available only electronically) we list the sources from this catalogue located within 5$^\prime$ from the maser positions. They are present in all cases except G 145.39+4.00. In Cols. (3)-(6) we give the angular offsets of the IRAS sources from the positions listed in Table 1, their coordinates and the IRAS names. In the following columns the flux densities in the four IRAS bands and their quality flags are presented.

Probably not all of these sources are physically associated with the CS cores. However, in all cases where we obtained the CS maps, there are IRAS point sources located within the regions of the CS emission (Sect. 3). We consider this spatial coincidence as a sufficient indication of the physical relationship. In the other cases, where the CS maps have not been obtained, there are IRAS sources within $\sim$1$^\prime$ from the maser positions. Most probably they are physically related to the CS cores.

2.4 Data reduction and analysis

  The data were reduced and analyzed in the same way as in Paper I. Briefly, the procedure was the following.

We have reduced the data and produced maps using the GAG (Groupe d'Astrophysique de Grenoble) software package. The measured spectra were fitted by one or more Gaussian components.

The next step in the data analysis was the derivation of the beam-averaged C³⁴S column densities when possible. We have performed it in the LTE approximation. The column densities were found from the C³⁴S line intensity integral assuming for the excitation temperature the value $T_{\rm ex}=10$ K (the dependence of the $N_{\rm L}$(CS) estimates on the assumed $T_{\rm ex}$ is rather weak; if $T_{\rm ex}$ increases to 20 K these estimates change by 35%).

We have calculated kinematic distances to our sources from the CS velocities according to the recommendations of Fich et al. (1989), i.e. for a flat rotation curve with $\Theta_0 = 220$ kms^-1 assuming a standard IAU value for R₀ of 8.5 kpc. For some sources no solution could be found in the framework of this model.

The sizes, LTE masses, virial masses and mean densities were calculated as in Paper I using either these kinematic distances or, when available, spectrophotometric distances to the exciting stars of nearby H II regions. For Gaussian brightness distribution our procedure gives the size at the ${\rm e}^{-1}$ level. As in Paper II we use "deconvolved" sizes, being defined as $\theta_{\rm deconv}=\sqrt{\theta^2-\theta_{\rm A}^2}$where $\theta_{\rm A}$ is the beam width at the same (${\rm e}^{-1}$) level. Of course, this procedure accounts for the beam size only approximately because many cores are elongated or have rather complex brightness distribution. The density estimates have been corrected accordingly.