From a Space.com article by Joe Rao.
I was wrong. My wife said last night that the moon was a blue moon. I said, that's not possible because a blue moon is a second full moon in the same month and that couldn't happen on 21 Nov. It has to happen at the very end of a month. She pointed me at the above article that explains it.
The short answer is that the two full moons in a month definition is actually erroneous.
Rao notes that Lawrence J. Lafleur, in Sky and Telescope in 1943, quotes a 1937 edition of the Maine Farmer's Almanac, stating that a blue moon is the rare occurance of four full moons in a season instead of three. The blue moon is the third moon in that season.
The mistake was apparently made by James Hugh Pruett in a 1946 Sky and Telescope where he misinterpreted Lafleur's explanation to mean a second full moon in a single month. (See the above article for details).
The wrong explanation was propagated by Deborah Byrd in the radio program “Stardate” in 1980 and then the wrong definition, quoting Rao, “went viral.”
It all makes me wonder when I first learned of the two-full-moons-in-a-month definition. I would have thought it was long before 1980, but maybe it was after that date.
Well it sort of spoils the whole thing. Now we'll have to talk about old-correct-definition blue moons and new-incorrect-definition blue moons and none of them will seem quite right any more.
This would all be a lot simpler if the moon would just actually turn blue once in a while.