In Ilkorin, nearly all original primitive final vowels vanished, which was also the rule in Noldorin. First all long final vowels were shortened, then the remaining short final vowels vanished. Any Ilkorin words ending in vowels were the result of some other process, such as final [j], [w] becoming [i], [u] (after which they became [e], [o] in most cases). The loss of final vowels in Ilkorin is very well attested in The Etymologies; see below for examples.

Final [-o] became [-a] after [g] and [ɣ]: One possible counter-example to this rule is ᴹ✶morókō > Ilk. broga (Ety/MORÓK). Here the final [-go] seems to have developed into [-ga] instead losing its final vowel, as suggested by Helge Fauskanger (AL-Ilkorin/broga). Another example (also suggested by Helge Fauskanger, AL-Ilkorin/-a) is the genitive suffix -a(n), which might have developed from primitive ᴹ✶ʒō [ɣō]; this example hints that this development occurred after [ɣ] as well. A third possible example is adda “daddy” (from ?✶attō), again suggested by Helge Fauskanger (AL-Ilkorin/adda). I find this last example less compelling, however, since this child’s word may simply be a late invention rather than a derivation from a primitive form. The theory presented here assumes that [-o] became [-a] after [g] and [ɣ] but not otherwise.

Final [-u] sometimes was preserved? Another possible preservation of a final vowel might be Ilk. ganu, whose Quenya cognate ᴹQ. hanu “a male” implies the survival of a primitive final [-u] (Ety/ƷAN). One possible explanation is that it was not a true cognate, but instead developed from a primitive adjective [ɣanwā], and thus had its final [u] develop from the [w], as suggested by Helge Fauskanger (AL-Doriathrin/ganu). The Noldorin development ᴹ√ƷAN > ON. anu > N. anw seems to contradict this scenario however, so perhaps the final -u was preserved by comparing the noun and adjective forms in both Ilkorin and Noldorin.

There are no clear examples of final [-u] being lost in Ilkorin. The only possible example is ᴹ✶orku [?urku] > urch (Ety/ÓROK), but in this case the short final [u] would have become [o] primitively, as indicated by its cognates ON. orko and ᴹQ. orko.