Annu. Rev. Astron. Astrophys. 1982. 20: 517-45 Copyright © 1982 by Annual Reviews. All rights reserved

2.2. CO Emission Surveys

In contrast to OH, CO emission is directly subject to the inverse square law, and so its detectability has distance limits imposed by beam dilution effects. However, where it is detectable, CO often can be mapped over fairly extended regions. A glance at Table 1 suffices to show the predominance of CO observations. Of course, observations within our own Galaxy have shown CO to be an excellent tracer of interstellar molecular gas (Burton 1976).

Although about 35 galaxies are now reported to contain CO (out of approximately 150 searched), questions of the relationship between molecular cloud content and other characteristics of a galaxy cannot be satisfactorily addressed until a statistically well-defined sample has been surveyed. Efforts to carry out such a survey are still being pursued. Rowan-Robinson et al. 1980) surveyed a large set of spiral galaxies, sacrificing line-profile information to obtain short integration times on the assumption that the global CO emission profile matches the 21-cm profile. Although they were able to show that galaxies with strong CO lines (T*_A > 0.2 K) are very rare, their several reported detections with T*_A < K 0.1 K indicated that a slight sensitivity improvement could yield different detection statistics. Verter (private communication) has begun a new survey using the high sensitivity and stable frequency response of the Bell Telephone Laboratories 7-m telescope.

One present complication in CO surveys is that investigators normally begin searching at the nuclear position. Depending on the relative angular sizes of the telescope beam and galaxy, this procedure leads to preferential detection of galaxies with central concentrations of CO. Few searchers continue looking away from the nucleus to find galaxies with CO emission exclusively from the disk, like M31. Another important effect is the dependence of beam dilution on inclination. Face-on galaxies (IC 342, for example) are most easily detectable because all the emission in the beam is concentrated within a narrow velocity interval (30-80 km s^-1). In contrast, the same emission from an edge-on galaxy is spread over hundreds of km s^-1, a velocity range that may be comparable to the irregularities in the spectral baseline.

An examination of the spiral galaxies known to contain CO reveals no correlation between CO abundance and any classical criterion such as morphological type or luminosity class (e.g. Morris & Lo 1978). The tendency of Rickard et al. (1977a) to detect late-type spirals appears to have been a selection effect. Barred spirals show no substantive difference from normal spirals in CO content (Elmegreen & Elmegreen 1982, Rickard et al., in preparation). At least for the spirals, it is still true that the distinctive features of galaxies with detected CO (dustiness, nuclear activity, etc.) also characterize some galaxies that currently betray no CO emission.

Elmegreen et al. (1980) sought CO emission at 52 positions toward 6 Magellanic-type irregulars, obtaining upper limits for the surface brightness significantly lower than the mean for nearby detected spirals. Israel et al. (1982) note a similar deficiency of CO within the Small Magellanic Cloud. In the Large Magellanic Cloud, however, CO emission has been detected in several HII regions (Huggins et al. 1975, Israel et al. 1982). It appears that, with the possible exception of the LMC, the Magellanic-type irregulars are CO deficient compared to our Galaxy. Elmegreen et al. (1980) offer two reasons for this: (a) a lower CO abundance in irregulars than in late-type spirals, or (b) a mean CO excitation temperature that is a factor of two or three lower than in our Galaxy, possibly reflecting a lower cosmic-ray heating rate. In galaxies of low metallicity (as at least some of the Magellanic irregulars are), a lower CO abundance could be accounted for by the increased CO formation time and by the increase in the photodissociation rate due to the relatively smaller amounts of obscuring dust.

Some elliptical galaxies have also been searched, on the grounds that mass loss from their red giants and supergiants could produce detectable amounts of CO (cf. Zuckerman 1980). To date, the only reported detection is a tentative one in NGC 185 (Johnson & Gottesman 1982). The associated H₂ mass is about 10⁴ M, which is 10% of the mass seen in HI.

Finally, we note the reports of identifications of uv transitions of H₂ and CO in the absorption spectra of high-redshift quasars (Varshalovich 1981), which indicate the presence of H₂ clouds in the disks of distant galaxies.