The values in this column must be 1, 2, вЂ¦, m where m = the number of replications. The data for Latin Squares with replication contains five columns with the first column representing the Replication. Once again the row and column values must be 1, 2, вЂ¦, r and the treatment values must be A, B, C, вЂ¦ until the rth capital Latin letter. The data for Latin Squares design without replication must contain four columns in the order: Rows, Columns, Treatments, Data. For Example 1 of Latin Squares Design, this format is shown in Figure 7В of that webpage. The input data may also be in standard (stacked) format. The row and column headings canвЂ™t be omitted and there can be no missing data values. In either case, the row and column labels must be numbers 1, 2, вЂ¦, rВ (where r = the number of treatments) and the treatment labels must be A, B, C, вЂ¦ until the r th capital Latin letter.

This last option is only available for Latin Square design without replications. For Example 1 of Latin Squares Design with the data as in Figure 4 of that webpage we would use the Excel format with data below treatments option. For Example 1 of Latin Squares with Replication with the data formatted as in Figure 1 of that webpage, we would use the Excel format with data to right of treatments option. You now fill in the dialog box that appears as shown in Figure 1.ДЇour input formats are accepted.

For example, to perform the analysis in Example 1 of Latin Squares Design with Replication, press Crtl-m, choose the Analysis of Variance option and then select the Latin Squares option.

The various capabilities described on the Latin Square webpages, with the exception of the missing data analysis, can be accessed using the Latin Squares Real Statistics data analysis tool.

R asreml latin square