The first number of the Empire Journal of Experimental Agriculture, of which this journal is the immediate successor, appeared in April 1933. It was part of a significant collaborative trend which included the Empire Marketing Board, the Commonwealth Agricultural Bureaux, and the International Wheat Pool, and led in time to the Commonwealth Secretariat and, indirectly, to FAO. The new journal was intended to serve agricultural progress in all the countries of the world which were connected with Britain, including South Africa and the Sudan. In the subsequent half-century of great change many new general and specialist journals have arisen, particularly in the more developed countries, in the Commonwealth and elsewhere; and this journal has come to be concerned mainly with production-oriented agricultural research in tropical, sub-tropical and winter-rainfall environments.

Most scientists now own, or have access to, an electronic calculator and appear to have absolute faith in the answers it gives. However, these calculators have design limitations which for some data sets will affect the accuracy of the calculations and so produce incorrect results. In statistics work, corrected sums of squares and products, needed for Analyses of Variance and Regression Analyses, can be affected. The first two sections of this paper describe these limitations and illustrate how they can affect the result of calculating the variance of a particular data set. A simple procedure is then described that will help overcome the problem. Finally, in anticipaion that the micro-computer will eventually replace the calculator as the main mathematical tool, a more general procedure is given for calculating the sums of squares and products.

A graphical method, based on bivariate analysis, is used to present yield data from intercropping experiments involving two crop species. The method is used to demonstrate two- and three-factor interactions in factorial experiments.

Three levels of N were imposed on mature Indian hybrid tea and green leaves were harvested according to different plucking standards at either 7 or 14-day intervals throughout the main production period. Larger rates of N and less selective plucking standards improved the fresh weight of shoots harvested but dry matter yields were less and the proportions of waste fibre in the finished product were increased. Thus, saleable black tea production was less responsive to N and plucking standard and the order of combined treatment effects was altered. Plucking interval did not affect yield but it changed the distribution of the crop in time. A 7-day interval gave the most uniform crop. These data are discussed in relation to the economics of tea production in Central Africa.

The black teas produced from the field treatments described previously (Cloughley et al., 1983) were evaluated chemically and organoleptically. Increasing N application rates and harvesting larger shoots had adverse effects on the liquor quality as assessed by theaflavin (TF) analysis, and on the value of the product as assessed by tea tasters' sensory examination. There were also large, regular intraseasonal fluctuations in TF content, small values corresponding to occasions of crop peaks. Shoots predominantly of 2 or 3 leaves and a bud harvested from plots given 135 kg N ha−1 generated the largest total cash value. A total value index (TVI), involving yield and quality, is introduced and illustrated.

The dependence of yields of sweet clover and alfalfa on environmental conditions, including soil acidity, at the germination stage has been studied over a period of 28 years. Results have been dramatic: sweet clover and alfalfa can grow well and produce large yields even on strongly acid soils (pHKCl 4.0) providing that the process of germination starts in a neutral medium.

In a 2 row cereal: 1 row pigeonpea intercropping system, four sorghum and two millet genotypes were examined with four of pigeonpea on a medium-deep Alfisol during 1978 and 1979. The cereals usually produced a large proportion of their sole crop yields. Millets matured relatively early which allowed large yields of later-maturing pigeonpea. A tall millet gave the best total Land Equivalent Ratio (LER) of 1.78 and most monetary returns. Early and/or short sorghums produced large LERs (1.51–1.59) and combinations with an early hybrid also gave good returns. A tall late sorghum gave poor yields of both components, smallest total LER (1.30), and little return. Pigeonpea intercrop yields became larger as the maturity difference between the cereal and pigeonpea increased. A combination of a short, early but large-yielding cereal with a pigeonpea that is as late as possible without incurring undue risk of moisture stress may be ideal.

Nodulation and N2-fixation of groundnut were investigated in sole and intercropping systems. Intercropping with pearl millet, maize, or sorghum reduced nodulation and N2-fixation. This effect is ascribed to the shading of groundnut by the cereal component and the consequent decrease in photosynthesis of the legume canopy.

Field experiments investigated the effects of seed rate, row spacing and time of harveston growth and yield of irrigated fibre flax and economics of fibre production at the Experimental Farm, Palampur, North Western Himalayas during two winter seasons between 1979and 1981. A seed rate of 60 kg ha−1 and row spacing of 20 cm gave the best yield and most profit. Any delay in harvesting after flower initiation significantlyincreased fibre yield and net profit per hectare.

Three levels of nitrogen (N) were applied to cotton grown in irrigated field trials at two locations in Zimbabwe in 1978. Dry matter (DM) production, total uptake and distribution of N among vegetative and reproductive components, and soil mineral-N contents were recorded about every 14 days. About 60% of total DM was produced, and 40% of total N taken up, between 10 and 16 weeks after sowing. Most N was present in vegetative parts, particularly leaves and branches, during early growth but, later, it accumulated in buds, flowers and bolls. At maturity, seeds and lint contained 42% of total above-ground plant N. N concentrations were similar in sympodial and mainstem leaves, petioles and branches. Inorganic N applied at sowing had little effect on plant N, but when given after 10 weeks it increased the N content of leaves, stems, branches, petioles and bolls.

The possibility of using seasonal changes in leaf nitrate reductase activity (NRA) as a reliable and sensitive indicator of plant nitrogen (N) status has been investigated in field-grown cotton. These changes were compared with those in nitrate concentration in petioles and variations in soil-N concentration. We conclude that NRA in the uppermost, fully-expanded sympodial leaves may provide a more convenient, sensitive and reliable indicator of plant-N status than measurements of nitrate concentrations in petioles.