Bullheading involves pumping produced fluids back into the formation using a kill-fluid. A key operational parameter is the required bullheading rate which depends on surface pressure, available horsepower, and erosion limits. There is wide variation in current guidelines for bullheading rates, especially for large-diameter wellbores. Therefore, a well-scale bullheading test program was conducted using a 5200-ft-deep vertical well with 9-5/8"x2-7/8" casing/tubing annulus located at LSU test well facility. The tubing was instrumented with 4 downhole pressure gauges and fiber optic DTS/DAS to obtain data on the downhole flow dynamics and determine bullheading efficiencies. In a typical test, a large nitrogen cap placed at the top of the annulus was bullheaded by pumping fluid in annulus with continuous returns taken from the tubing side. Tests were conducted with varying fluid rates (50 to 500 gpm), initial gas-cap size (30-60 bbl), gas pressurization method and kill fluids (water and synthetic base mud). It was observed that the bullheading process involves simultaneous gas compression, gas bubble breakage, gas dispersion, and gas displacement, unlike the typical assumption of bullheading a large gas slug. The breakage of the initial gas slug depended on the surface pressure and the extent of gas-liquid mixing. The minimum water flowrate required for gas bullheading matched to water velocity just above small bubble velocity in water. Increase in water flowrate increased the bullheading efficiency, e.g., bullheading with 350 gpm required <50% water volume compared to 150 gpm water flowrate. Experiments with a highly pressurized initial gas cap and a larger initial gas cap volume resulted in relatively more efficient bullheading due to lower slip velocity resulting from higher average gas-holdup in the gas-swarm. In one test, the gas was bullheaded for some time and then allowed to migrate upward in a shut-in well. It was observed that the gas migration velocity (0.71 ft/sec) was higher than the gas slip velocity during bullheading (0.3-0.6 ft/sec). Contrary to the popular belief, the gas also did not carry its pressure while migrating in a shut in well. The experimental observation of bubbly flow instead of slug flow during bullheading under sufficiently higher surface pressure helped understand the multiphase flow dynamics of bullheading and it can help provide realistic bullheading guidelines based on well conditions.