In both Sindarin and Noldorin, its seems a final [wi] or [ui] became [y], which resulted several abnormal plural forms. This development is most obvious in Noldorin examples from The Etymologies:

• ON. malui > mely, plural of (h)mâl “pollen” (Ety/SMAL).
• ON. orkui > †yrchy, plural of orch “Orc” (EtyAC/ÓROK).
• ON. rankui > †rhengy, plural of rhanc “arm” (Ety/RAK).

The last two plurals were marked archaic, reformed to more ordinary plurals yrch and rhenc, respectively; (h)mâl also had a more normal [Noldorin-style] plural meil. In these words, the ancient final u was short (compare ᴹQ. malo, orko, and ranko) and thus the ancient plural had diphthongal -ŭi. In the Noldorin plurals, the i survived long enough to cause i-affection, or perhaps the final y itself also produced i-affection.

There is also a Sindarin example of a similar change:

• S. i·chery, plural of herw [heru] “wine”, derived from primitive ✶syeru or [sy]erwā (PE23/139).

This Sindarin example is not entirely analogous to the Noldorin examples, since Tolkien explicitly stated that the Sindarin word herw was derived from herwā [sic., actual primitive would be *syerwā]. Thus the plural development would be syerwāi > χerwī > hery. If this distinction is significant, than in Sindarin the development of final y may be limited to words ending in w after final vowel losses, and possibly also ancient long ū (we have no examples). Unlike Noldorin, it seems i-affection was at least partially inhibited, since we see hery and not **hiry; perhaps i-raising was prevented by the archaic final diphthong -wī or -ūi.