Howard Clase schreef in
I think it's because precedence takes precedence even over poor Latin.
I.e. whatever name was given in the first "legal" description stands; even
if there's a mistake.

*****
That is not quite it. For example if the Original Publication would use an
epithet "arboricolus" this would be correctable to "arboricola" no matter
what the gender of the genus (Art 23.5).

Decisions on correction of spelling tend to be involved and sometimes
controversial.
*****

Linneus himself made a similar mistake
when naming the Cat's Ears as Hypochaeris rather than Hypochoeris, since
the Greek word for pig he based it on is choeros. He later corrected
himself, but the rules say that since he used an -a- the first time this
spelling must stand. (You will still find both spellings in use in modern
books - about 50:50. This makes me sound very erudite, but in fact this was
discussed in these very columns a few years ago and I am only repeating what
came out then!)

Howard Clase