1994A&A...289..357D

We have used the Australia Telescope Compact Array to survey 21-cm absorption toward compact continuum sources in and behind the Large Magellanic Cloud. We find many absorption lines: out of a sample of 30 lines of sight observed we detect absorption in 19. This shows that the cool atomic phase of the interstellar medium is abundant in the LMC, more so relative to the warm neutral medium than in the Milky Way. This abundance of cool atomic clouds contrasts with the relative scarcity of molecular clouds in the LMC, suggesting a difference in the cloud population compared with what is found in the Milky Way. An alternative interpretation which is consistent with our results is that the temperature of the cool atomic clouds is much lower in the LMC than in the Milky Way, perhaps 40K compared to 60K in the solar neighborhood, but with a similar abundance. Further observations will resolve this ambiguity. The spatial distribution of the absorption suggests that the abundance of cool-phase gas in the LMC decreases with distance away from the region of intense star formation in the vicinity of 30 Dor at the north end of the giant molecular cloud. Within about 1kpc of 30 Dor all lines of sight show cool gas, whereas beyond about 6kpc none do. This region of intense star formation and abundant molecular gas around and south of 30 Dor weighs so heavily in the cool gas statistics that it accounts for all the difference between the LMC and the Milky Way. Considering only our lines of sight that do not pass within one kpc of 30 Dor, we find exactly the same distribution of warm and cool H I phases as in the solar neighborhood. Most of the absorption lines detected are quite narrow, with half power widths of only 2 to 3km/s. They thus allow us to probe the kinematics of the cloud population in detail. We find random (one-dimensional) velocity dispersion of about 12 km/s relative to the large scale velocity field as mapped in 21-cm emission surveys. Some of this velocity dispersion may be due to the finite thickness of the H I distribution coupled with the velocity gradient along each line of sight through the LMC. Assuming cylindrical rotation for the disk, the low velocity H I emission, the so called "L" component of Luks & Rohlfs (1992), must be located on the far side of the LMC disk. The strong star formation activity close to and south of 30 Dor has to be even further away on the back side of the "L" component.