On pg. 112 I mention how speakers of Latin sometimes interpreted the Greek accent in Greek loanword as Latin stress, which brought about exceptional stress assignments like Phílippus. I should also have mentioned that Quintilian (Inst. 1.5.22) implies that Camillus and Cethēgus could be stressed on the antepenult:
Cum acuta et gravis alia pro alia ponuntur, ut in hoc "Camillus", si acuitur prima, aut gravis pro flexa, ut "Cethegus".
When acute and grave are switched as in the case of Camillus, if the first syllable is given an acute, or when a grave is substituted for a circumflex as in the case of Cethegus.
These exceptional stress patterns are probably reproducing the original Etruscan initial-syllable stress pattern since both forms are thought to be borrowed from Etruscan.