We have reported that contrary to theories of Evidence Integration (EI), time in visual perceptual decision-making is needed not only for accumulating external sensory evidence, but also for collecting samples from static internally represented distributions as it is predicted by the theory of Probabilistic Sampling (PS) (Cosyne 2013; VSS 2013). However, the precise relationship between EI and PS in the general case of dynamic stimuli is still unknown. We used an estimation-based variant of the classical random dot motion task, where in each trial, participants reported their best estimates of stimulus direction and their subjective uncertainty about it. Across trials, we varied the strength of the sensory evidence and the trial time, and across experiments, we varied the trial sequence volatility (intermixed versus blocked coherence levels). We found a marked decrease in error-uncertainty correlation within the first 300-500ms of the trial, indicating EI, followed by a significant increase throughout the trial, indicating PS. Importantly, the transition between these segments shifted as a function of signal coherence and volatility. Consequently, EI and PS during decision making work in parallel, with EI taking the lead early but PS dominating the later part of the probabilistic process.