A process of allocating treatments to the experimental units so that each experimental unit has an equal chance of receiving any of the treatments. It is a necessary condition for obtaining a valid estimate of the experimental error. Location of treatments by chance so as to provide unbiased estimates of treatment means and experimental error.