Please Note: the e-mail address(es) and any external links in this paper were correct when it was written in 1995, but may no longer be valid.
Onsala Space Observatory, S-43992 Onsala, SWEDEN
The Large Magellanic Cloud (LMC), the nearest neighbour of the Milky Way, is a small, irregular galaxy at a distance of 50 kpc. LMC provides a hostile environment for molecular gas. The environment differs from that in the Galaxy in several aspects. Relative to Galaxy, LMC exhibits a lower metallicity, stronger radiation fields and a higher gas-to-dust ratio. This implies less shielding and higher photodissociation rates and is expected to lead to smaller and more clumpy clouds, lower molecular and dust abundances, higher temperatures and lower optical depths for the molecular gas. The proximity of the LMC enables detailed studies of the molecular gas in an environment differing from that in quiescent spiral galaxies such as the Milky Way.
The study of molecular clouds in the Galaxy has resulted in a number of scaling relations between global cloud properties such as mass, density, luminosity, size and velocity dispersion (e.g. Larson (1981),Scoville & Sanders (1987),Solomon et al. (1987),Myers & Goodman (1988a),Myers & Goodman (1988b)). Of these quantities the luminosity and the velocity dispersion are well defined and directly measurable. On the other hand cloud mass cannot be measured directly and the cloud size has to be defined. Attempts have been made to ``explain'' the scalings in terms of virial equilibrium, turbulence, equipartition of energies (Larson (1981),Solomon et al. (1987),Chieze (1987),Fleck (1988),Myers & Goodman (1988a),Myers & Goodman (1988b),Henriksen (1991)), but so far no explanation is fully satisfactorily.
Star formation is believed to take place mainly inside molecular clouds. Thus, it is important to be able to estimate the masses of molecular clouds. However, molecular clouds consist mainly of molecular hydrogen, but being a symmetric molecule it lacks radiation at suitable frequencies. Therefore, masses are estimated by relating the intensity () of a ``tracer'' to the column density () of molecular hydrogen using an empirical ``conversion factor'' :
Such tracers are, for example, other molecules (notably CO and its isotopomers), dust and gamma-rays.
The CO to conversion factor has been calibrated through observations of molecular clouds in the Galactic molecular ring (e.g. Scoville & Sanders (1987),Solomon & Barrett (1991),Wolfendale (1991)). X is usually taken as constant and frequently used to estimate molecular gas masses in external galaxies. The constancy of X is questionable; both theoretical considerations (e.g. Kutner & Leung (1985),Maloney & Black, (1988),Elmegreen (1989)) and observations of the outer parts of the Galaxy (Sodroski (1991)) and irregular galaxies (e.g. Israel et al. (1986),Cohen et al. (1988),Johansson (1991),Rubio et al. (1993)) indicate that X depends on the environment.
Observations of the CO transition in the 30 Dor and N 159 regions in the LMC were made within the SEST Key project ``CO in the Magellanic Clouds'' using the SEST (Swedish-ESO Submillimeter Telescope) at La Silla, Chile. The FWHP beam width at 115 GHz is , which corresponds to about 10 pc at the distance of the LMC.
The CO emission in the 30 Dor and N 159 regions can be decomposed into 25 and 13 individual clouds, respectively. The cloud radii range up to 23 pc (which is considerably less than for the largest cloud complexes found in the Galaxy, 50 to 100 pc). As mass estimate we use the so called virial mass (which follows from the simplest form of the virial theorem, see e.g. MacLaren et al. (1988)):
where is a coefficient depending on the density distribution, is a measure of the cloud size (``radius'') and is the one-dimensional velocity dispersion.
If is expressed in , in parsecs, in and a radially symmetric density profile is used, then . We define the cloud radius as , where and are the spatial dispersions from a two-dimensional Gaussian fit to the integrated intensity contour map of the cloud. The velocity dispersion is determined from a Gaussian fit to the global line profile of the cloud.
Figure 1: Log-log scaling of the line width versus the size.
Figure 2: Log-log scaling of the luminosity versus the size.
Figure 3: Log-log scaling of the virial mass versus the luminosity.
Scatterplots of the resulting scalings are shown in Figures 1, 2 and3. The filled circles are for the N 159 region, the empty circles are for the 30 Dor region and the line is a fit for Galactic clouds taken from Solomon et al. (1987). For the N 159 clouds good correlations are found for the different quantities. Data for the 30 Dor clouds show a larger scatter, which is probably due to lower signal-to-noise ratios and difficulties in defining cloud sizes. The scatter plots indicate (despite of the large scatter in the 30 Dor data) that the same line width-size relation holds in 30 Dor, N 159 and Galactic clouds. The luminosity-size and virial mass-luminosity scalings are however different for 30 Dor and N 159. A clear offset is evident in the plots. This suggests that the line width-size relation is fundamental and that differences in the physical conditions of the clouds and their environments play a role for the two other relations. Maloney (1990) showed that the virial mass-luminosity relation likely is an artifact of the line width-size relation. Our data support this suggestion. Further more, our study shows that the clouds in the 30 Dor region are underluminous compared with those in the N 159 area. Differences in optical depths and/or volume filling of CO gas may provide an explanation.