In this study, a 10-yr (1998-2007) climatology of observations from the Tropical Rainfall Measuring Mission (TRMM) satellite is used to study regional mechanisms of monsoon onset across tropical and subtropical South America. The approach is to contrast regional differences in the structure, intensity, and rainfall of mesoscale convective systems (MCSs) prior to and after onset, in the context of thermodynamic conditions from the National Centers for Environmental Prediction (NCEP) reanalysis data. This is accomplished by analyzing the mean annual cycle time series, 10-yr frequency histograms, and 3-month-averaged values prior to and following onset in four regions of distinct rainfall variability. Observed MCS metrics and NCEP variables include lightning flash rate, convective rain fraction, height of the 30-dBZ isosurface, minimum 85-GHz polarization corrected temperature, and the fluxes of sensible and latent heat. The west-central Amazon region had a distinct maximum of MCS intensity 2 months prior to the monsoon onset date of each region, which was well correlated with surface sensible heat flux, despite the observation that thermodynamic instability was greatest after onset. At the mouth of the Amazon, the dry season rainfall minimum, the premonsoon maximum in MCS intensity metrics, and monsoon onset were all delayed by 2-3 months relative to the west-central Amazon. This delay in the annual cycle and comparatively large difference in pre- versus postonset MCSs, combined with previous work, suggest that the slow migration of the Atlantic Ocean intertropical convergence zone controls onset characteristics at the mouth of the Amazon. All metrics of convective intensity in the tropical regions decreased significantly following onset. These results, in the context of previous studies, are consistent with the hypothesis that thermodynamic, land surface, and aerosol controls on MCS intensity operate in concert with each other to control the evolution of precipitation system structure from the dry season to the wet season. The other two regions [the South Atlantic convergence zone (SACZ) and the south], associated with the well-documented dipole of intraseasonal rain variability, have a weaker and more variable annual cycle of all MCS metrics. This is likely related to the strong influence of baroclinic circulations and frontal systems in those regions. In the south, fewer but larger and more electrified MCSs prior to onset transition to more, smaller, and less electrified MCSs after onset, consistent with previous climatologies of strong springtime mesoscale convective complexes in that region.