As a follow-up to the discussion on Talat Chaudhri's facebook wall, I'd like to sum up where we are standing on the issue.
We have KK <Nadelik>, UC/R <Nadelek>, RLC <Nadelak/Nadelik> and the following attestations collected by Gendall (1992 & 2008):
TC, DP, and myself agree that **Nadelik looks like a Lhyudian Cymricism and that the native LC form more likely ended in -ak (which stood for [-ək] or [-ak], according to theory).
But what about the first syllable? Is it feasible that i-affection had spread, and that the underlying MC form would have been /ne'delek/ (c.f. Breton Nedeleg)? Or is /na'delek/ more likely?