Problem with na in nls

I'm trying to run a nls on a subset of a data.frame.
In the subset, one observation is NA. So I drop the
observation but when I ask for :

sm  <-  nls(machin$revcum ~

Lc.singh(machin$popcum,p), start=list(p=c(2,3)))

I get :

Erreur dans parse(file, n, text, prompt) : syntax
error in "~ "

If I put some value for the non available observation
instead of droping it, it works. So what's the problem ?