Tamhane's T2 verses Dunnett's C and T3 Post Hoc test

I have been working really hard to find the answer to this question, but the statistical references I have read all say they won't go into detail about these techniques. They give a citation, but I don't have access to it. I have asked the statistics professor in my department, but he doesn't really work with ANOVA too much and wasn't able to help.

So here is my question: Which post hoc one-way ANOVA test, Tamhane's T2, Dunnett's C and Dunnett's T3, is better under what circumstances?

The Games-Howell test is too liberal for my situation as the sample sizes are less than twice as different between my groups.

Any help in this matter would be greatly appreciated!