Molecular Clouds are the birthing grounds from which nascent stars emerge. Whilst a paradigm for isolated low-mass star formation (SF) has been developed - the "Standard Model" (Shu, Adams & Lizano 1987, ARA&A, 25, 23) - this model's applicability to massive SF is uncertain at present. Indeed most low-mass stars actually form in clustered regions of high-mass SF (e.g Clarke et al. 2000, Protostars and Planets IV, 151).

Massive protostars produce strong winds, outflows and a considerable UV photon flux - these phenomena will drastically alter the physical conditions and structure of the surrounding natal cloud material. Two possible outcomes of these complex feedback effects are envisaged:-

• Further induced SF, initiated by the compression of molecular material.
• Suppression of SF, due to cloud disruption.

The current generation of SF models (e.g. Bate et al. 2002, MNRAS, 332, L65-L68) are tackling SF within the context of a dynamic clustered environment. Surveys of Giant Molecular Clouds (GMCs), the sites of cluster formation, will provide the observational constraints for these theoretical models.

We have recently completed data reduction for a ~3000 arcmin² survey of the W3 GMC at 850 microns using SCUBA's scan-mapping mode. W3 is located in the Perseus spiral arm at a distance of 2.4 kpc and comprises some 10⁵ solar masses of molecular cloud material spread over a 60 pc region (Lada et al. 1978, ApJ, 226, L39). It contains a number of young HII regions and newly formed complexes of massive stars; W3 Main, W3 North, W3 (OH) and AFGL 333. At 850 microns SCUBA's resolution is 14'' (FWHM) - corresponding to 0.16 pc at W3's distance.

Data analysis is in the early stages, however considerable clump and filamentary structure is present in the vicinity of the IRAS-traced star formation regions located along the NE edge of the cloud, adjacent to the W4 HII region. The SF within this region is believed to have been triggered by the expansion of the HII region (Lada 1978). Figure 1 shows the mapped region overlaid with CO J=1-0 emission line contours from the FCRAO Outer Galaxy Survey (Heyer et al. 1998, ApJS, 115, 241) (Click here for full-size image). The contours represent the corrected antenna temperatures in the velocity range -57 to -32 km/s, contour levels are 12, 56, 100 and 144 K. Since the large scale structure in SCUBA scan-maps is unreliable it has been removed by convolving the original map with a Gaussian beam of FWHM 136'' (twice the maximum chop throw), and then subtracting the convolved map from the original. The resulting map was then convolved with a Gaussian point spread function (14" FWHM). There is a good correspondence between the highest contours of CO emission and the brightest continuum emission. Figure 2 (click here for full-size image) presents the SCUBA map with the positions of the main SF regions indicated. The blue crosses represent the positions of the nine 0-type stars which comprise OCl352, the star cluster which is ionizing the W4 HII region (Normandeau et al. 1996, Nature, 380, 687-689).

Within the region surveyed we have identified some 250 objects, using the Williams clump-finding algorithm (CLFIND2D - Williams et al. 1994, ApJ, 428, 693). We have determined the clump positions and sizes and are in the process of estimating clump masses using temperature determinations from IRAS HIRES data. Ongoing work includes:-

• Determination of the local clump-mass spectrum - these results will be compared to theoretical predictions of the mass-fragment spectrum in collapsing compressed shells.
• Evaluation of clump number density as a function of cloud position. This will reveal the extent to which clump structure is formed preferentially close to known SF sites as part of the process of induced SF.
• Combination of the W3 SCUBA data with our unbiased CO J=1-0 survey of W3 (taken at FCRAO 14m with beam-width, 44'' at 115 GHZ, sampling) and HI data from the Canadian Galactic Plane Survey. This will enable us to study the cloud kinematics, e.g. the bulk gas motions within the cloud and whether the cores detected at 850 microns exhibit high velocity gas (typically associated with molecular outflows).

The analysis of the W3 dust continuum emission, coupled with cloud kinematical data (CO and HI) will provide a powerful set of constraints which models of SF will need to reproduce.