1. Two experiments were conducted in a lowtemperature cabinet. In the first experiment three grass varieties (two North African fescues and S 22 Italian rye-grass) were subjected to two cutting treatments (herbage cut from 8 in. to 2 in. and from 4 in. to 1 in.). Dry-matter production, dry-matter content and tiller number were recorded from plots of 9, 16 or 25 plants. The second experiment, a uniformity trial, was planted with a uniform population of a North African fescue. The optimum plot size and shape within the cabinet, and the optimum arrangement of blocks, was determined using a Titan autocode programme.

2. Treatment and variety differences could be detected using very small plots. For the material and apparatus used, an 18-plant plot was the optimum size.