Two rice trials were conducted from 2005 to 2006 in rice research institute, Kala Shah Kako Pakistan to evaluate the efficiency of alpha lattice design in field experiments. The average standard error of difference between genotypes mean is used to calculate relative efficiency of alpha lattice design. Both experiments clearly identified the advantages of small blocks. The average gain in efficiency was 119% with maximum 128%. Mean ranks comparison for both randomized complete block and alpha lattice design were performed. It was observed that the ranks were not constant across the experiments. The results emphasize that the traditional randomized complete block designs (RCBD) should be replaced by alpha lattice in the agricultural field experiments when number of varieties to be tested in an experiment increases to more than five or ten. In such a situation finding a homogeneous block is quite difficult in field experiments.