2014MNRAS.443L..49P

The relation between the star formation rate (SFR) and the stellar mass of star-forming galaxies has been used to argue that major mergers cannot be the main driver of star formation. Here, we re-examine these arguments using the representative IMAGES-CDFS sample of star-forming galaxies at z = 0.4-0.75, taking advantage of their previously established classification into pre-fusion, fusion, and relaxing galaxy mergers. Contrary to previous claims, we show there is no tension between the main-sequence scatter and the average duration of the fusion SFR peak. We confirm previous estimates of the fraction of SFR due to morphologically selected galaxies ( ∼ 23 per cent) or the SFR enhancement due to major merger during the fusion phase ( ∼ 10 per cent). However, galaxy mergers are not instantaneous processes, which implies that the total fraction of the SFR associated with galaxies undergoing major mergers must account for the three merger phases. When doing so, galaxies involved in major mergers are found to represent 53-88 per cent of the total SFR at z ∼ 0.6. The fraction of luminous infrared galaxies (LIRGs) in the fusion phase is found to be in agreement with the observed morphological fraction of LIRGs without discs and with the observed and expected major merger rates at z ≤ 1.5.