As rightly noted, the DMRT is a widely criticised statistical posthoc test especially in soil science. This is because it has much less control over the Type I error rate. If one needs to necessarily do mean separations, then the Holm-Sidak or the Student-Neuman-Keuls method is a much better option

Please sir which soft ware can this (Holm-Sidak or the Student-Neuman-Keuls method) be found?

I think R, SAS, SIgmaPlot and sometimes excel can do these tests. Which kind of datastet do you have and which software is available...because other methods like Tukey HSD is also a viable option. See here for some alternative options for analyses: http://www.real-statistics.com/one-way-analysis-of-variance-anova/unplanned-comparisons/

Thank you Sir.

Dear Arthur, Unfortunately i cant assess this link. can you send the soft ware setup to me though: [email protected]

The link that was sent was for information, it was not a software link. Sorry I cannot send you a software setup primarily due to copyright reasons. Which software do you use?

Presently am using GenStat, SPSS, and MegaStat

I think its possible to do similar test using these software....check this out for SPSS:

http://egret.psychol.cam.ac.uk/statistics/local_copies_of_sources_Cardinal_and_Aitken_ANOVA/MultipleComparisons_3.htm

Thanks am going to look at the content