THE TULLY-FISHER RELATION FOR THE CO LINE

The question addressed here is whether linewidths measured in CO emission can be used in the Tully-Fisher relation as a substitute or supplement to H i line widths. To answer this question we have observed a test sample of galaxies in the Coma-A1367 supercluster using the IRAM 30 m telescope to measure CO line widths for comparison with 21 cm line widths from the Arecibo telescope. The galaxies were selected to span a very broad range of H I line widths (113-575 km s-1) and to have simple, two-homed line shapes at 21 cm. For those detections with good signal-to-noise ratio we find a nearly perfect correlation between line widths in CO and H I. The dispersion in the ratio of the two line widths is only 0.09. Thus there is no question that CO line widths will correlate well with absolute magnitude as in the Tully-Fisher relation. The main issue is resolution: even at the Coma distance (60 h-1 Mpc) the IRAM beam resolves the CO distribution in some of these galaxies, and it is useful to combine the signals from several beam areas to improve the correlation in line widths. The line shapes in CO are very different from the corresponding H I line shapes, which is a result of the difference in spatial distributions of the two species. Evidently the CO extends beyond the peak of the rotation curves of most spirals, since there is such a good agreement between widths measured with the two tracers. If anything, the CO lines are consistently slightly broader than the H I lines. Using model line profiles we investigate the effect of different radial distributions of gas on line shapes and widths. Eventually we may hope to use CO to improve on distance estimates using the Tully-Fisher relation, and to extend its use to more distant galaxies and clusters than H I telescopes can reach. More extensive investigations will be needed to distinguish which tracer gives better distance precision.