Plant populations, nitrogen levels and the use of alpha naphthalene acetic acid (ANA) to rein-force natural flower differentiation were compared in a field experiment at Malkerns in the Swaziland Middleveld. Yield was raised by increasing plant population above 17,000 plants per acre, but reached a maximum at 23,000 plants per acre. Increased yield from denser planting occurred at the expense of fruit size and fruit set, and of sucker and slip production. Hormone counteracted the depressing effect of denser planting on fruit set and denser planting reduced sunburn damage and raised fruit acidity. 400 lb N per acre per crop increased yield over 200 lb N by increasing fruit size in the mother crop, and by increasing fruit size and set in the ratoon. Added nitrogen markedly reduced fruit acidity. ANA improved fruit set and therefore yield, without detrimental effects on fruit shape or quality. There are grounds for small increases in population above 17,000 plants per acre, for the use of ANA to reinforce natural flower differentiation and for the cautious use of the higher level of nitrogen fertilization.

Herbage production from twelve perennial grasses was analysed under two frequencies of defoliation. There were differences in the speed of establishment from seed, Heteropogon contortus, Pennisetum pedicellatum and P. polystachyon being most rapid. Once the grasses were established, the highest yields were obtained from Andropogon gayanus and Brachiaria brizantha and the lowest from Paspalum notatum. Yields were higher at the lower frequency of cutting. There were large differences in production throughout each year, with yields during winter and the dry season (November to April) only about one-tenth of those obtained in the rainy season. A. gayanus, B. brizantha, Chloris gayana and Dichanthium caricosum gave higher yields than other species during winter. Variations in yield between species could generally be ascribed to variations in the extension of leaf area, although some species maintained high net assimilation rates during unfavourable growth periods.

The perennial grass Pennisetum polystachyon was grown alone or in mixture with the indigenous legumes Atylosia scarabaeoidesand Clitoria ternata and the exotic species Calopogonium mucunoides, Centrosema pubescens and Stylosanthes gracilis. Only the last two persisted for more than one or two years. Mean annual yield from the mixture containing Centrosema was 61 to 65 q./ha., of which 20 per cent was legume herbage. The protein content of grass was about 3 to 6 per cent and of legumes 10 to 14 per cent. Mixed swards produced as much protein as in a previous experiment where pure swards of Pennisetum polystachyon received 44. kg./ha. of nitrogen in fertilizer form. Soil nitrogen increased under the Centrosema mixture and yields from the mixture increased during the experiment. Yields of grass herbage from the mixtures were generally lower than those of the grass swards. The reduction in yield of grass grown in mixed swards was mostly due to decreases in the area of foliage rather than to reductions in net assimilation rate.

The analysis of a uniformity trial on cassava is briefly discussed, mentioning possible effects of plant losses on the methods of analysis. Two methods of calculating comparable variances, on a per plant and per unit area basis, gave very high values for the index of heterogeneity, with a mean of 0·94. When the ratio of overall cost per plot to cost per unit area lies between one and two, an optimum plot size of between 16 and 32 plants is obtained. The comparable plot sizes would range from 192 to 385 sq. ft and give a coefficient of variation less than twenty.

Experiments in bulbil nurseries and field sisal were carried out in N.E. Tanganyika, 5°S, 39°E, 200 m. above sea level, at a mean annual rainfall of 1160 mm., on deep red loam. The discussion covers the establishment of sisal through a polythene mulch, the control of weed growth, the persistence of polythene of different thicknesses and from different sources, and some effects on growth of the sisal and on soil properties. A mulch of black polythene improved the growth of bulbil nurseries by 30 per cent or more, due to conservation of soil moisture and elimination of weed competition and disturbance associated with the removal of weeds by hoeing. For nurseries 150-gauge material was preferred, but there were important differences in the persistence of material purchased from two sources. Where black polythene is used in a nursery the density can be increased from 80,000 to 110,000 plants per hectare, to help offset the high cost of the material. A mulch of 250-g. black polythene controlled weeds in field sisal but did not improve growth.

Field experiments in three successive growing seasons (1963–66) studied effects of variety and plant density on yield components of broad beans. In 1963/64 total seed yield of a local variety (Beladi), increased with increasing seed rates from 70 to 280 lb per feddan, but yield increment diminished beyond 140 lb. The same results were obtained with Beladi and Rebaya 34 in 1964/65 and 1965/66, but Rebaya 34, yielded less than Beladi at all seed rates. Growth analysis revealed that Rebaya 34 had greater growth rate at the beginning of the season, due to its relatively large seed size, but eventually Beladi overtook it. Maximum pod growth in both varieties occurred when leaf dry weight was declining. The proportion of dry matter was greater in pods of Beladi than Rebaya 34 because of more sinks in the former variety. The present seed rate of 70 lb per feddan seems to be suboptimal and factors affecting establishment and/or effective leaf area at the time of pod growth are detrimental to the final yield.

The results are reported of pot experiments, using indicator plants, on the nutrient status of three soils varying in fertilizer history. All soils were deficient in nitrogen, and in the absence of added nitrogen yield of grass was below 10 per cent of the potential. All forms of nitrogenous fertilizers were equally effective. All soils showed an acute deficiency of phosphorus at the early stages of the experiments, but with time, and depending on the amount of fertilizer applied in the past, the response to phosphorus decreased. There was little or no response to potassium at the early stages of the experiments, but relative yields later fell linearly with time. Calcium was necessary in small quantities as Ca(OH)2, CaCO3 to correct soil acidity, particularly for legume growth. Magnesium increased yields at the latter stages of the experiments. Sulphur was deficient in all soils. The soils had adequate amounts of the minor nutrients. A cheap and quick technique was used with simple and factorial designs to obtain preliminary information, as a guide to planning detailed field experiments on the nutrient requirements of perennial crops like coconuts and other intercrops.

A procedure is described in which variation in the attenuation of β-radiation by water held in nylon pads is used to show changes in the water content of soil. The method enables sequential measurements to be made in narrow zones of soil under field conditions; the results are not influenced by the presence of soluble salts.

Component analysis was applied to the interpretation of (a) leaf analyses from six similar groundnut fertilizer trials, and (b) the chemical composition of different parts (root, stem and successive leaves) of the sugar cane plant. For each of the groundnut trials it was found that approximately 60 per cent of the total variation in N, P, K, and Ca could be accounted for by two independent linear functions of the elements, representing the balance of N and K with Ca and the balance of N with P and K. The second of these functions was consistently affected by phosphorus applications. Yield was related to one or other function, according to trial, but never to them both. Throughout the sugar cane plant the elements N, P, K, Ca, and Mg were found to be distributed between the roots and the aerial parts, and between active photosynthetic and non-photosynthetic tissues. In both cases the result was a reasonable and consistent simplification of a considerable body of data, such as had not resulted from a series of univariate analyses of variance.