There seems to be solid evidence that Latin letter names were indeclinable.
But in Greek, several letters' names do fit into standard declension patterns: sigma, for instance, might actually be a -ma noun from sizō "to hiss". And gamma and digamma look like -ma nouns, even if they historically aren't.
So: do we ever find sigmata or gammata? Or, less obviously, alphai and betai? Or are all letter names indeclinable in Greek?
EDIT: I'd also be interested in cases of omicra and epsila, even though those names are later. But first and foremost I'm curious about usage in the classical period.