Photoluminescence blinking--random switching between states of high (ON) and low (OFF) emissivities--is a universal property of molecular emitters found in dyes, polymers, biological molecules and artificial nanostructures such as nanocrystal quantum dots, carbon nanotubes and nanowires. For the past 15 years, colloidal nanocrystals have been used as a model system to study this phenomenon. The occurrence of OFF periods in nanocrystal emission has been commonly attributed to the presence of an additional charge, which leads to photoluminescence quenching by non-radiative recombination (the Auger mechanism). However, this 'charging' model was recently challenged in several reports. Here we report time-resolved photoluminescence studies of individual nanocrystal quantum dots performed while electrochemically controlling the degree of their charging, with the goal of clarifying the role of charging in blinking. We find that two distinct types of blinking are possible: conventional (A-type) blinking due to charging and discharging of the nanocrystal core, in which lower photoluminescence intensities correlate with shorter photoluminescence lifetimes; and a second sort (B-type), in which large changes in the emission intensity are not accompanied by significant changes in emission dynamics. We attribute B-type blinking to charge fluctuations in the electron-accepting surface sites. When unoccupied, these sites intercept 'hot' electrons before they relax into emitting core states. Both blinking mechanisms can be electrochemically controlled and completely suppressed by application of an appropriate potential.