AbstractIn this article, we present the first‐time observations of binary raindrop collisions in rainfall events. These observations constitute a critical step in concluding a long‐standing debate on the controlling physical process, binary raindrop collision versus spontaneous raindrop breakup, for the raindrop size distribution (DSD) evolution from cloud to ground level. Our raindrop collision observations were made possible by a new instrument called the High‐speed Optical Disdrometer (HOD) that we recently developed for precipitation microphysics investigations. Our approximately 1 year long field campaign that covered 33 rainfall events provided 11 observations of binary raindrop collisions and outcomes but no spontaneous breakup observation. The field‐observed collision rate (i.e., number of raindrop collisions within the measurement volume of the HOD per unit time) showed an increasing trend with increasing rain rate as expected from the theoretical collision rate predictions. Furthermore, the field‐observed collision rates were (i) comparable (for rain rates less than approximately 50 mm/h) and (ii) significantly larger (for larger rain rates) than the theoretically predicted rates that have been used in various numerical investigations that suggest the controlling role of raindrop collisions in DSD evolution. Our observations, yet to be supplemented with observations from comprehensive field campaigns at different geographic locations and rainfall events for a definitive conclusion, support the collision‐driven DSD evolution hypothesis.