4 The CO data

Figure 2: Left panel a): total molecular gas as a function of blue luminosity. Right panel b): the same as in the left panel. Additional samples of ultraluminous infrared galaxies, galaxies in clusters, and spiral galaxies are included. Symbols are the same as in Fig. 1. Luminosity is in $L_{\odot}$ and mass is in $M_{\odot}$.

Millimetric observations were carried out at the Swedish-ESO (SEST) 15 m radiotelescope at La Silla in October 1996 and September 1998 during good weather conditions. In the first run we used the SESIS 100 receiver with a 1 GHz bandwidth at 115 GHz (¹²CO(1-0)). Typical system temperatures were $\sim$250 K (in the $T_{\rm A}^{*}$ scale) at the elevation of the sources and typical zenith opacities between 0.1-0.2. During the second run we used the IRAM 115 and IRAM 230 receivers with 500 MHz and 1 GHz bandwith, at 115 GHz (¹²CO(1-0)) and 230 GHz (¹²CO(2-1)), respectively. The half power beamwidth of the SEST at 115 GHz is 45'' and 23'' at 230 GHz.

All galaxies were observed at the central optical coordinate. Integration times were 2-3 hours depending on the signal-to-noise achieved. The pointing was regularly checked on nearby SiO masers. The pointing uncertainties were of the order of 5''. CO emission was detected in 47 galaxies and had low signal-to-noise detection in only 5 galaxies, ESO-LV1080110 (HDS), ESO-LV1880170 (CS), ESO-LV2850050 (HDS), ESO-LV3550300 (CS), and ESO-LV6050070 (CS). We have not included these galaxies in our analysis.

Two galaxies, ESO-LV3470340 and ESO-LV4060250, were considerably larger than the SEST beam and were observed in 5 and 7 positions, respectively, spaced by half of a beamwidth (23''). In Appendix A we show each position along the major axis of the galaxy and give their spectra. We have added the intensities at each position in order to obtain the total CO intensity of each galaxy.

The CO spectra were reduced with the CLASS package (Forveille et al. 1990). We have binned the spectra with a boxcar function. Spectra were corrected for first order baseline in most of the cases or third order in a few obvious cases where first order did not give a good fit to the data. CO intensities were calculated by using the main-beam efficiency, $\eta_{\rm mb}$, values of 0.7 and 0.5 for 115 GHz and 230 GHz, respectively. We estimated the 1 $\sigma$ uncertainty in the integrated line intensity taking into account the channel-to-channel noise (rms), the width of the emission profile ($\Delta V$) and the number of channels (N) that the emission profile covers ( ${\rm error = rms}\times\Delta V\times N^{-1/2}$).

Table 3 lists the CO data as follows. Column 1: designation in the ESO-Uppsala catalog (LV89); Col. 2: type of sample (control $\rm sample=CS$ and high density $\rm sample=HDS$) and morphological type (LV89) $\rm 1=Sa$, $\rm 2=Sa$-b, $\rm 3=Sb$, $\rm 4=Sb$-c, $\rm 5=S$..., $\rm 6=Sc$, Sc-d, $\rm 7=S$../Irr, $\rm 8=Sd$; Col. 3: velocity derived from central CO(1-0) profiles in kms^-1; Col. 4: the width of the emission profile in kms^-1; Col. 5: blue luminosity in $L_{\odot}$ derived from $B_{\rm T}$ magnitude (errors in $L_{\rm B}$ are within 10% when the magnitude estimates in the RC3 have errors of 0.1 mag); Col. 6: far-infrared luminosity in $L_{\odot}$ calculated as described in the next section; Col. 7: CO intensity in the line J=(1-0) in K kms^-1 and errors; Col. 8: H$_{\rm 2}$ masses and errors in $M_{\odot}$ estimated from the velocity integrated CO(1-0) emission as described in the next section, and Col. 9: CO intensity in the line J=(2-1) in K kms^-1. Distances were corrected for the Virgocentric flow according to model 3.1 in Aaronson et al. (1982). Hubble constant value of 75 kms^-1 Mpc^-1 was adopted in all calculations.

Table 4 lists the CO intensity in the line J=(1-0) available in the literature for 7 galaxies (4 in the HDS and 3 in the CS). The differences between the fluxes we have measured and the ones obtained previously are due to (i) different sizes of the beam (Elfhag et al. 1996); (ii) baseline adjustments (Combes et al. 1994; Andreani et al. 1995), or short integration time (Horellou & Booth 1997).