MANCOVA mixed model by sas

[water]

I meet a MANCOVA mixed model by sas of question.
Here, fruit_set and seed_set as two dependent variables, with pollination_treatment as fixed effect, year as a random factor, with plot_density as covariate. I could not to write this sas program. Could you help me to do it, thanks advance to you!
In fact, because fruit_set(ratio of fruit to flowers per plant) and seed_set (ratio of seed to ovules per flower) may be intercorrelated, so I have to select MANCOVA.
If just one dependent variable, for example just fruit set, I write program by proc mixed,
-----------------------
proc mixed data = abc;
class trt year;
model fruit_set = trt plot_density trt*plot_density;
random year year*trt year*trt*plot_density;
run;
-----------------------
or two dependent variables, but not a random effect, items of 'year' as a fixed effect, I write program by proc glm,
------------------------
proc glm data = abc;
class trt year;
model fruit_set seed_set = trt | year | plot_density;
manova h =_all_;
run;
-----------------------
I try to write "MANCOVA mixed model" by proc mixed, but could not run in SAS!
----------------------
proc mixed data=abc;
class trt year;
model frut_set seed_set = trt | plot_density;
random year year*trt year*trt*plot_density;
run;
----------------------
Please help to revise my program "MANCOVA mixed model" or rewrite a new program, thanks advance!!!

[lgwaters]

I have exactly the same problem. Does anyone have a solution to this for SAS?

[Rick_SAS]

See if this blog post "MANOVA or mixed model?" answers your question. It mentions PROC MIXED  (REPEATED / SUBJECT=subject)

and how to specify hypothesis tests that are analogoud to GLM's MANOVA statement.

If not, please post your GLM code.

[lvm]

There are also previous discussions on this Stat Procs community board on this subject.