Because editors of the Nag Hammadi texts differ so widely in their practice, it seems best to state the principles on which the present Coptic edition is based.

The character of our textual witnesses demands a cautious approach. For four tractates our sole witness is a Cairo manuscript, CG II; for the other two (tractates 2 and 5) it is joined by insignificant fragments of parallel manuscripts. The evidence is thus virtually limited to a unique codex, with the result that critical editing must proceed by conjecture.1 Moreover, the archaic date and orthographic peculiarities of ou