On pg. 392–3 in discussing the archaic 3rd sg. perf. ending -eit I give the form FVEIT from ILLRP 918 ( = CIL I
2.1297). But, in fact, Degrassi prints FVVEIT. Unfortunately the inscription from Rome is now lost. According to Degrassi only Marini, an 18th century scholar, read FVVEIT, while others have FVEIT. Since FVVEIT is the lectio difficilior, it is probably what the inscription had.
There are a number of other inscriptional attestations of the perfect ending -eit, which should be mentioned: POSEDEIT (ILLRP 517, Isoverde, probably composed at Rome, 117 BCE), PROBAVEIT (ILLRP 379, Rome, 62 BCE beside COERAVIT). As a matter of fact none of these is old enough to prove an original diphthong *ei. At most they confirm the evidence of Plautine scansion that one variant of the 3rd singular perfect ending inherited a bimoraic nucleus before the final consonant. Another set of inscriptional 3rd sg. perf. forms with the spelling -EIT consists of cases where -EIT is preceded by I (mainly compounds of īre 'to go' like REDIEIT (ILLRP 122, Rome, mid 2nd cent. BCE), VENIEIT (Lex Agraria, ll. 58, 65, 67, 75, 91, 111 BCE, and OBIEIT (ILLRP 589, Ferentium, Modern Ferento, 67 BCE, according to the reading of A. Emiliozzi 1983:701; AE 1980:0371, Casa Biagi; CIL 10.1935, 14 BCE), but also POSIEIT 'placed' (AE 1987:0252, Sulmo)). In these forms other factors may be involved, viz. the avoidance of graphic II and/or a dissimilation of ii to ie (Kent 1912 and Weiss OHCGL:429). In any case, although Latin inscriptional evidence does not strictly prove the diphthongal origin of the ending -EIT, there is no other plausible source.
The originator of the idea that -eit is from recharacterized -ei appears to be J. Vendryes, who proposed this solution in 1937.
Emiliozzi, Adriana. 1983. Sull'epitaffio del 67 a.c. nel sepolcro
dei Salvii a Ferento. MEFRA 95.2:701–17.
Kent, Roland. 1912. Dissimilative Writings for ii and iii in Latin.
Transactions and Proceedings of the American Philological Association 43:35–56.
Vendryes, J. 1937. La 3e personne du singulier du perfectum latin.
Revue des étude indo-européennes 1:3–5. Reprinted in Vendryes, J. 1952.
Choix d'études linguistiques et celtiques. Paris: Klincksieck, 156–8.