Santo Domingo, in addition to being two words is also a recognized phrase (recognized, that is by the Controlled Vocabulary). When you enter a CV phrase, it's treated as a single understood term, and the search engine will return images that are tagged with that term. When you put an "AND" between Santo and Domingo, the search engine assumes you want images that contain those two terms ("Santo" and "Domingo") and not the single understood CV term "Santo Domingo".
I hope that made it more clear, but I suspect I only made it more complicated.