data alfcut;
  infile 'alfalfacut.txt' firstobs = 2;
  input block variety $ date $ yield;
  
proc mixed method = type3;
   class block variety date;
   model yield =  block variety date variety*date / ddfm = kr;
 
 /* KR and Satterthwaite adjustments are equivalent for */
 /*   balanced split plot and similar for split plots */
 /*   KR necessary for more complicated repeated measures */
 /*  use /ddfm = satterth;  to get the Satterthwaite adj. */
   
   random block*variety;
   
   lsmeans variety date variety*date /diff;
   
   estimate 'ladak - ranger' variety 0 1 -1;
   title 'Split plot';
   
run;