Stellar masses for thousands of z>1 resolved, dusty starbursts

Dusty star-forming galaxies, with individual star formation rates >100Msun/yr, formed most of the stars in the Universe at early epochs (z 1-2), yet their physical origins and triggering, whether it be from major mergers or secular disk-bound star-formation, is still unsolved. In assessing the role of major galaxy mergers (amongst dusty galaxies) to cosmic star formation, the measurement of galaxies' stellar masses is critical. Starbursts' stellar masses tell us about their past average star formation rate and whether or not the current high-SFR phase is indeed rare (short-lived) or in line with expectation. Can we definitively measure whether the occurrence of major mergers at a given star-formation rate varies with stellar mass, as predicted? Or are high star formation rates simply always indicative of short-lived bursts? Here we propose deep infrared imaging of a new 1.77deg^2 extragalactic legacy field which has very unique, deep, high-resolution radio interferometric coverage from the e-MERLIN SuperCLASS survey; the dataset's radio continuum mapping will allow a unique morphological measurement of obscured star formation in 5000 starbursts, probing the merger fraction of dusty galaxies (which are too dusty for HST morphology). Critically, Spitzer IRAC 3.6um and 4.5um will allow calculation of galaxies' stellar masses and allow us to directly test if higher mass starbursts are less likely to exhibit clumpy, merging morphologies than their lower mass counterparts (a key corollary to the galaxies `main sequence' framework). In addition, IRAC coverage provides an independent AGN indicator (steep mid-IR powerlaw), will improve the quality of photometric redshifts in the field, and will provide crucial near-IR positional counterparts for future submillimeter coverage.