[R] MANOVA with within-subject factors

[rab]

Hi. I have a split-plot design where one of the treatment factors is 
between-subjects and the other two factors  are applied within each 
subject in 2x2 design. In addition, the subjects occur in blocks, 2 to a 
block.

In a between-subjects design with repeated measures, it's easy to see 
how to set up the formula. The response matrix consists of columns, each 
column a response variable (or repeated measure at each time point with 
each column a time point). But I don't see how to do the multivariate 
repeated measures analysis with within-subjects factors. The univariate 
analysis is:

my.aov <- aov(y~f*g*d + Error(block/f),data=sm2)

where f is the between-subjects factor and g and d are within-subjects 
factors.

I believe this can also be analyzed as a multivariate design but I don't 
see how to set it up.

Rick B.