I was able to see this effect while reproducing a classic physics experiment to measure the ratio of specific heats of a gas. The idea was to send a measured amount of energy into a wire filament suspended in the center of a sealed volume of air and measure the resulting pressure rise for which I’d built a tiny pressure gauge. Observing the fluid level in this gauge through a microscope, I would discharge a precision capacitor across the filament. The gauge fluid would quickly rise, hold constant for a second or so, and then drop back. What was happening was that, as the filament heated the air in contact with it, the pressure would rise as the heated air expanded. It would stay constant as that small volume of heated air rose like a helium-filled balloon by convection toward the top of the sealed volume, touching nothing but the air around it. When that heated air contacted the cold surface (in this case, a gallon jug in a water bath), it lost heat to it and the pressure dropped.