On the chronology and the phonetics of the change o > u in final closed syllables in Latin

It has generally been assumed that the Old Latin vowel weakening o > u in final closed syllables occurred before word‑final m and word‑final s at the same time. The present article argues that this assumption is incorrect, since the inscriptional evidence clearly shows that the change before m took place later than before s. This chronological difference is due to the different phonetic properties of ‑om [‑õ] and ‑os [‑os]. It is also suggested that, at least in some cases, the sound change ‑om > ‑um could have had analogical support as well