Still, the lower case σ,ς are more lunate than not...
In this blog at the British Library you may monitor the transition from manuscript to print (Aldus Manutius) in late 15th century. The Florentine Homer manuscript (1466) of the British library still uses C.
In the first Aldine Greek book in print (1495),¹ however, you see Σ,σ, and final ς (an evident lunate holdover) supplanting their lunate analogs you'd see in contemporary manuscripts; lower-case σ, a less evident lunate morph, features routinely in byzantine manuscripts as well, e.g. 11th century, even 9th century; a Greek minuscule development expert might be able to track its history.
I am no expert to know the reasons Aldus Manutius had the print font Σ "restore" the non-lunate form, but it looks like once he got the ball rolling, the print world never looked back. This is the system the West inherited; you might be amused by Newton's mid-17th century notebooks.
¹ It appears to replicate an earlier, first printed book entirely in Greek, "Erotemata or Grammatica" of Constantine Lascaris (1434–1501), produced in Milan in 1476, in the printing house of Diogini da Paravicino (Dionysius Parvisinus). It still uses the lunate C!