(Apologies for opening a new thread - I intended to read the list only online but now I see that if I do that I cannot hit reply…)

Thanks, Matt. So I suppose that means it doesn't happen only on my machine.

Though I notice it isn't round-tripping correctly either, e.g. the iotas are not being re-subscripted... I suppose that's a limitation of Unicode itself, though.

Correct - that's actually (for me) the intended behaviour since it allows me easily to convert subscript iotas to adscript ones.

But the important thing is that it's not limited to iota subscripts. Take the following two words from the text quoted in the earlier post: ἀλλοῖα αὐτὸϲ

Capitalized: ἈΛΛΟΙ͂Α ΑΥ̓Τ

Decapitalized: ἀλλοῖα αὐτ

These are not isolated cases. For example, any text from the Thesaurus Linguae Graecae (www.tlg.uci.edu/demoinfo/demoauthors.php) will do.

Take Matt's credo, as it appears on the website: Πάντες ἄνθρωποι τοῦ εἰδέναι ὀρέγονται φύσει. σημεῖον δ’ (21) ἡ τῶν αἰσθήσεων ἀγάπησις· καὶ γὰρ χωρὶς τῆς χρείας ἀγαπῶνται δι’ αὑτάς, καὶ μάλιστα τῶν ἄλλων ἡ διὰ τῶν ὀμμάτων. οὐ γὰρ μόνον ἵνα πράττωμεν ἀλλὰ καὶ μηθὲν μέλλοντες πράττειν τὸ ὁρᾶν αἱρούμεθα ἀντὶ πάντων ὡς εἰπεῖν (25) τῶν ἄλλων.

Capitalization, followed by decapitalization, results in a net loss of two words and one period - and there were no iota subscripts: πάντες ἄνθρωποι τοῦ εἰδέναι ὀρέγονται φύσει. σημεῖον δ’ (21) ἡ τῶν αἰσθήσεων ἀγάπησις· καὶ γὰρ χωρὶς τῆς χρείας ἀγαπῶνται δι’ αὑτάς, καὶ μάλιστα τῶν ἄλλων ἡ διὰ τῶν ὀμμάτων. οὐ γὰρ μόνον ἵνα πράττωμεν ἀλλὰ καὶ μηθὲν μέλλοντες πράττειν τὸ ὁρᾶν αἱρούμεθα ἀντὶ πάντων ὡς εἰπεῖν (25)