On pg. 467 I mention the word ōvum 'egg' which seems not to have been affected by whatever rule changed *oktōwos into octāvus. This may have been due to the ō occurring in the initial, i.e. in the Proto-Italic stressed syllable, as I said in the book, or, as Alexis Manaster Ramer suggests to me, due to the absolute initial position of the ō. In any case AMR also calls my attention to the strange fact that the Romance languages reflect Proto-Romance open o (Ital. uovo, Sp. huevo, Fr. oeuf), which normally is the reflex of a Latin short o. There is no absolutely satisfactory explanation for this. Meyer-Lübke suggested ōvum became ōum by regular loss of w before a back vowel and that became oum by pre-vocalic shortening. The w was then restored from the genitive ōvī. Rohlfs also starts from ōum but since the reflexes of long ō and short u would both have been a close o he suggest that the first of the two identical vowels was dissimilated to an open o. This seems a bit more straightforward and has the parallel of Ital. tuo < *tuoo, cf. the plural tuoi, reflecting an open and not the expected close o in the first syllable. In any case the long vowel of Latin ōvum is very well established starting from Ennius' Ova parire solet genus pennis condecoratum.