Joerg van den Hoff <j.van_den_hoff <at> fzd.de> writes: > > > res <- nls( y ~ (x * Bmax) / (x + Kd), start = list(Bmax=, Kd=)) > > (providing some sensible start values for the free parameters , of course) > > cf. `nls' manpage for details There are also "self-starting" models -- SSmicmen I think? Maybe in nlme package? Ben Bolker