Paul Derouda wrote:But as I'm sure Annis didn't take βροτός as an example by chance, I'm left wondering if there's always an explanation of this kind...
Quite a few instances of this are the result of historical changes of the sort you mention here, syllabic resonants in proto-Greek resolving into their modern forms. Another common situation is a name that would otherwise be unusable, such as Ἀφροδίτη (the iota is long), before Κρονίων, etc.
I'm sure some German scholar in the 1800s wrote a monograph giving every instance of this, but I don't have access to such a tool.
Seymour goes into a little detail on this, Monro (§370) rather more.