From a recent post:
"This is because air dissolved in the liquid will tend to come out of solution at low pressures, and contribute a partial pressure of air to the contents of any macroscopic cavitation bubble. When that bubble is convected into a region of higher pressure and the vapor condenses, this leaves a small air bubble that only redissolves very slowly, if at all."
I previously posted a description of what may be an analogous event which I encountered when I had a small mining operation. It was evidently disregarded as unworthy of comment
. It may be that I do not understand the above, but it seems to imply a similar phenomenon.
To amplify my original post slightly I'd just add that the gradual accumulation of air in the pump casing had little or no noticeable effect until it suddenly reached a critical threshold, at which point the pump quite abruptly ceased to deliver. This ocurred several times, after repeated re-priming, before the cause of the problem (increased restriction of flow in the suction line) dawned on me.