Examples are taken from three biometrical texts, two unpublished experiments and Yates (1937) to illustrate quantitatively how the lay-out of the plots can affect the estimate of error variance in field experiments and how confounding of treatment and positional effects can arise. In two examples in which the plots did not run the whole length or width of the block, the estimates of error were reduced by one half and one quarter respectively by allowing for sub-blocks within blocks or for rows and columns. In split-plot experiments where the main plots were split breadthwise, the sub-plot error was about the same size as the main-plot error or larger whereas it is expected to be at most one third that of the main plots when these are split lengthwise. Serious bias can arise in the estimate of treatment effects due to confounding with positional effects; this is illustrated.

On-farm experiments were used for assessing the processing and food value of four promising sorghum cultivars and the willingness of farmers to pay for the herbicide 2,4-D for controlling the parasitic weed Striga hermonlhica in coarse grains. The active participation of farmers in these experiments was useful in determining the conditions under which the improved sorghum cultivars and the herbicide would be most likely to he adopted by smallholder farmers in southern Mali.

Ten weighable containers, the same size as 4-gallon petrol tins, were located in a plot 80 metres from the leading edge of a 2-acre field of irrigated alfalfa growing under advective conditions in the Sudan. Evapo-transpiration from alfalfa growing in these containers was compared with evaporation from a series of small evaporimeters level with the top of the plants and from a standard U.S. Class A evaporation pan, and also with estimates of evaporation calculated by the Penman formula. Evaporation from the free water surfaces in evaporimeters or pan was always greater than from alfalfa in the containers when the crop was 25 cm. high or less, but lower when the crop was more than that height, whilst the figures calculated by the Penman formula were much lower than either. Cumulative water losses were also recorded from containers where the soil was kept wet or allowed to dry out, and these were compared with evaporimeter losses up to the stage where no further water was being lost from the dry containers.

The proliferation of statistical methods and software increases the need for authors to state, unambiguously and informatively, the methods they have used. The reader of any paper should expect to learn precisely what the author has done, and to be able to adopt similar procedures in his or her own work.

The effects of irrigation on the yields of wheat and cotton were studied on a sandy clay loam soil, with a water-table at 7 ft below ground, under the arid conditions of Lyallpur. The conventional practice of providing 19-acre in. of water to Mexican wheat and 25-acre in. to American cotton proved to be a wasteful and unproductive practice on such soils. High yields of both these crops could be obtained without any subsequent watering after an initial 4-in. irrigation applied for seedbed preparation and sowing, while one 3-in. irrigation after sowing was needed to obtain maximum yields.

The effect on cross pollination of four directions and six distances from the marker variety in Egyptian cotton was investigated. The percentage of cross pollination did not differ appreciably in the different directions but there were significant differences in cross fertilization among the six distances. A linear relation was observed between cross pollination per cent and distance up to 8.8 m. from the marker variety. Implications of these findings on various aspects of cotton breeding methodology are discussed.