The boiling liquid expanding vapour explosion (BLEVE) is well known but not well understood. Some still argue about what comes first, the BLEVE or the vessel rupture. Some believe the BLEVE is triggered by some pressure transient inside the vessel and this causes a superheat limit explosion which causes the vessel to rupture. Others believe it is the vessel rupture by some weakening process that leads to the BLEVE. This paper will provide evidence that the latter description that is correct for most, if not all BLEVEs observed in practice. This paper describes small scale experiments of aluminum tubes that were weakened by machining a thinned wall area over a specified length. The tubes were filled to a desired level with liquid propane and then the propane was uniformly heated electrically until the tubes failed. The failure pressures ranged from 10 to 33 bar. The tube was instrumented to capture failure characteristics (pressure, temperature) and consequences: blast overpressure and imaging of the propane cloud and shock around the vessel; ground force under it; transient pressure and imaging of the boiling process inside the vessel. The work was done to improve our understanding of the fluid – structure interactions during the fire heat induced failure of a pressure vessel holding a pressure liquefied gas. We were specifically interested in the near field hazards including blast overpressure and ground force. This paper will focus on the early milliseconds of the process where the vessel begins to open and a shock wave is formed and moves out into the surroundings. The imaging reveals presence of a Mach shock at the exit of the vessel at the early stage of the opening. A chronology of the event also shows that the lead shock is generated early in the explosion process, and is long gone before the liquid starts boiling, arguing that vapour expansion is the main contributor to the first shock overpressure.