6 Outline of main results

  Our sample of IRAS sources contains objects with a wide range of bolometric luminosity and we have therefore examined our data for correlations between the various quantities measured. When doing this, one should bear in mind the fact that the different molecular tracers sample gas of greatly differing properties. For example, the measured linewidths show that the HCN(1-0) and HCO⁺(1-0) lines are in large part formed in the outflows. The compactness of the methyl cyanide emission together with the high excitation temperatures shows that CH₃CN traces dense gas close to the IRAS source. At least in the case of IRAS20126+4104 (Cesaroni et al. 1997a), there is evidence that the methyl cyanide emission comes from a disk like structure at distances of around 1000 AU from the young star.

The effectiveness of HCO⁺ and HCN as outflow indicators is demonstrated in Fig. 18 where we have plotted the FWZI of these lines against the corresponding widths for CO taken from Table 2. Also shown is a comparison between the FWZIs of HCN and HCO⁺ (bottom panel). One sees that in most sources HCO⁺ and HCN, which preferentially sample high density gas, have higher FWZI than CO. One should be clear that this is partially an artifact. The CO data taken from the literature have in general been averaged over larger beams than those ($\sim$25$^{\prime\prime}$) used for HCN and HCO⁺. Thus, they sample gas farther and at lower velocity from the IRAS source. Moreover, the FWZI used in this comparison is a quantity which depends on S/N and hence is not necessarily a good measure of source characteristics. Nevertheless, the line widths measured in HCO⁺, HCN are extremely high and show that a non-negligible fraction of dense gas is being accelerated to velocities above 50 km s^-1. It is interesting to note also that there is no clear correlation between the FWZI measured in HCO⁺ and HCN, and bolometric luminosity (see second panel of Fig. 18).

  

Table 10: Results of Gaussian fits to CH₃CN lines towards centre position in map

^a Fixed.

  

Table 11: Source FWHP and deconvolved angular diameter (in parentheses) in different lines

^a Detected only towards H₂O maser position.

We do however find a correlation between the intensity ratio of the HCN(1-0) and HCO⁺(1-0) lines and the dust temperature as measured by the ratio of 60 $\mu$m and 12 $\mu$m IRAS fluxes (see Fig. 19 top). We note here that this flux ratio sometimes is found to be less a measure of temperature than a measure of the prevalence of small grains excited by ultra-violet photons and out of thermal equilibrium. We do not know whether that is the case for the sources in our sample. However, we clearly see in Fig. 19 top, tendency for larger I(HCO⁺)/I(HCN) and hence presumably a larger [HCO⁺]/[HCN] abundance ratio in sources with lower apparent dust temperature (higher 60/12 flux ratio). The correlation is not of great statistical significance, but seems to us worthy of verification with a larger sample of objects. We note moreover that there is no clear correlation with bolometric luminosity (Fig. 19 bottom) but there is a slight correlation (Fig. 19 middle) when one plots I(HCO⁺)/I(HCN) against I(CH₃CN)/$F_{100~\mu{\rm m}}$. The motivation for considering the latter quantity is that it may be a measure of the relative importance of a possible disk in these sources (assuming that the methyl cyanide emission in some sense measures the disk emission). Here one sees a slight fall-off in the HCO⁺ to HCN intensity ratio with increasing relative importance of the methyl cyanide emission. Methyl cyanide seems to appear preferentially in high temperature regions and hence it is possible that Fig. 19 top and middle both are showing a correlation between high dust temperature and high [HCN]/[HCO⁺] abundance ratio (or alternatively between the fraction of "very small grains'' and the [HCN]/[HCO⁺] abundance ratio).

  

Table 12: Rotation temperatures and column densities derived from the rotational diagrams of CH₃OH (see Fig. 11)

^a Source averaged, assuming the deconvolved CH₃OH angular diameters given in Table 11.