The first number required on the Variance header line is the number of sections in the data. A typical experiment has 1 section. However, when combining results from several experiments into one analysis, we would have a section for each experiment. Then each experiment will have an error structure with its own variance. While there may be correlations among residuals within a section, are independent across sections.

The number may be given as the name of a factor with the appropriate number of levels as in the example below ( site 3).


The following variance structure lines relate to an analysis three variety trials with common entries. It is assumed the data is sorted by experiment and that there is a complete grid of plots for each experiment although the experimental layout (number of rows and columns) may differ.
 met test run
  col 15     # Actually 12 12 and 15 for the sites respectively
  row 34     # Actually 34 34 and 28 for the sites respectively
  chks 7
  test 336
  geno 337
  yld 1  !*.01
  site 3
 metS0.dat   !section site
 yld ~ site ,
  !r at(site,3).row .02 at(site).col .90 .40 .036 site.test
  site 2 1
 12 col AR .1271    !S2=2.19
 34 row AR .751
 12 col AR .25      !S2=0.84
 34 row AR .56
 15 col ID          !S2=0.19
 28 row AR .38
 site.test 2
 site 0 CORUH 0.3
  .2 .1 .1

Return to start