DETERMINING STRUCTURE IN MOLECULAR CLOUDS

We describe an automatic, objective routine for analyzing the clumpy s tructure in a spectral line position-position-velocity data cube. The algorithm works by first contouring the data at a multiple of the rms noise of the observations, then searches for peaks of emission which l ocate the clumps, and then follows them down to lower intensities. No a priori clump profile is assumed. By creating simulated data, we test the performance of the algorithm and show that a contour map most acc urately depicts internal structure at a contouring interval equal to t wice the rms noise of the map. Blending of clump emission leads to sma ll errrors in mass and size determinations and in severe cases can res ult in a number of clumps being misidentified as a single unit, flatte ning the measured clump mass spectrum. The algorithm is applied to two real data sets as an example of its use. The Rosette molecular cloud is a ''typical'' star-forming cloud, but in the Maddalena molecular cl oud high-mass star formation is completely absent. Comparison of the t wo clump lists generated by the algorithm show that on a one-to-one ba sis the clumps in the star-forming cloud have higher peak temperatures , higher average densities, and are more gravitationally bound than in the non-star-forming cloud. Collective properties of the clumps, such as temperature-size-line-width-mass relations appear very similar, ho wever. Contrary to the initial results reported in a previous paper (W illiams & Blitz 1993), we find that the current, more thoroughly teste d analysis finds no significant difference in the clump mass spectrum of the two clouds.