R help - May 2010 - GAM vs. GAMM, how to model increasing error variance?

I fit a GAM to turtle growth data using mgcv:
>m1 <- gam(growth~s(mean.size,
bs="cr")+s(year,bs="cr",k=7)+s(cohort,bs="cr")+s(age,bs="cr"),
data=grow,
family=quasi(link="identity"))


The errors are skewed (and seem to be correlated with age) (code and plots
here:
https://docs.google.com/fileview?id=0B1cQ7z9xYFl2OWY2MzNlYzUtMWZmMS00YjExLTk0MjMtMGU3MmRhYzI5N2Vm&hl=en).


I tried fitting a oversimplified GAMM,
>mm1 <- gamm(growth~s(mean.size,
bs="cr")+s(year,bs="cr",k=7)+s(cohort,bs="cr")+s(age,bs="cr"),
data=grow,
family=quasi(link="identity"))

which dealt with some of the correlation... but... I don't know what/how to
specify a structure for the errors.


Any help/suggestions greatly appreciated,
Thank you,
Claudia

	[[alternative HTML version deleted]]