R help - Jun 2009 - Comparing nonlinear regressions for different subsets of data

I have a dataset with four different treatments and two different values
for each sample. I need to compare the subsets(the different treatments)
somehow. The data look suspiciously much like some kind of e-function to
me, or maybe michaelis-menten so they are not linear. With linear models
it's straight forward enough but what about nonlinear regressions? With
linear model I do:
data<-read.csv(file, header=T)
attach(data)
mod1<-lm(column1~column2)
mod2<-lm(column1~column2+treatment)
anova(mod1,mod2)

The I can also go on to get the interaction by:

mod3<-lm(column1~column2*treatment)

But the same kind of methodology does not seem to work for nls
functions. How should I do? Can I linearize the data? how to do that in
a good way? The best would of course be if I could do it without
linearization.
Regards
Martin