Abstract
Data are studied from 56 similar maize fertilizer trials from a range of dissimilar sites in the West Indian with a view to comparing way of allowing for the effects of environ-mental difference within the area of an experiment. The standard method is the use of blocks and it was sometime very successful, especially when there was a lot of variation to control. However, used inappropriately it can do more harm than good. Another method is adjustment by neighbouring plots, which on the average was even more successful though it too could do harm. A third method is the use of row-and-column designs. It was best of all when used well but it can do even more harm if used inappropriately. Also it can be very restrictive in its design considerations. Recommendations are made to deal with practical cases.
