So you want to control the pattern of randomization across subjects? So, e.g., if you have 100 subjects, then 25 will have seen 1a, 25 will have seen 1b, 25 will have seen 1c and 25 will have seen 1d?