1. Early Roman times, its popularity increases significantly. Already by the midfirst century ce, ephebic texts exhibit the ῑ → ει interchange more frequently than the `standard' convention of writing etymologically long ῑ with ι. In the second century ce, ει becomes the overwhelming orthographic norm for ῑ, being attested roughly twice as frequently as ι for ῑ. After 200 ce, ῑ → ει comes to be attested about ten times more frequently than ι for ῑ. This convention is frequently attested in lexemes like μειλήσιος, τειμ-, νεικ-, ἐλευσείνιος, ὁπλείτης, ἀφροδείσιος, and names derived from Ἶσῐς (e. g., εἰσίδωρος). There are just a few cases of ῐ → ει in the first century ce. Most examples of this interchange are from the second century ce and after, though they are not nearly as common as ῑ → ει during the same period (Threatte;In Attic Koine
2. The relevant fragments are collected, translated and discussed by Chahoud 2019: 51-53. Lucilius' orthographic prescriptions are contrasted by him with a different
3. viving fragment of his that argues for the practice (Dangel 1990, Chahoud 2019:
4. 63-65). Under the Accian scheme, long vowels are written double, hence ⟨PAA