6 Comparison with other surveys

Figure 8 shows the antenna temperature of the CS(2-1) line in the survey of Bronfman et al. (1996) plotted against the same quantity in our survey. The survey of Bronfman et al. was carried out also using the 20-m Onsala radio telescope. The straight line is the least-squares linear fit to the data. Its equation is:

y=(0.91 $\pm$ 0.16)x+(0.16 $\pm$ 0.19)

and the correlation coefficient is r=0.85. Since both data sets have been obtained from the same line using the same instrument, ideally the slope should be unity and the offset at x=0 should be zero. The statistical errors of these two quantities (1$\sigma$) are such that they allow these ideal values. The dispersion of the measured values from the fitted line is consistent with a calibration uncertainty of typically 20-30%. This is expected for an observing site at sea level such as Onsala.

Figure 9 shows the antenna temperature of the CS(1-0) line in the survey of Anglada et al. (1996) plotted against the antenna temperature of the CS(2-1) line in our survey. The survey of Anglada et al. was carried out using the 37-m radio telescope at the Haystack Observatory. The HPBW of this telescope is 41'' at the CS(1-0) frequency, i.e. similar to the HPBW of the Onsala telescope at twice the frequency. The least-squares fit to the data is described by:

y=(0.17 $\pm$ 0.02)x+(0.27 $\pm$ 0.12)

and the correlation coefficient is r=0.95. Since both the (2-1) and the (1-0) lines are optically thick, the ratio between the antenna temperatures in Anglada's survey and in ours could be explained by the ratio between the beam efficiencies of the Haystack telescope (9%) and of the Onsala telescope (56%): 0.16.

There seems to be a systematic offset between the antenna temperatures measured by Anglada et al. and our values by $\sim 0.3$ K. The reason for this is unclear. However, the statistical significance of this offset is poor ($\sim2 \sigma$).