Abstract
It is suggested that there is a great need for more practical experimentation, which can only be brought about by the closest co-operation between the engineer and the statistician. Emphasis is also placed on the value of cost-benefit analysis both for the actual research investment and in any resultant suggestions for improved design.