The partitioning of uptake of soil nitrogen between legumes and grasses in mown, mixed swards was studied at two sites in south-eastern Queensland. The swards contained either Lotononis bainesii, Desmodium intortum or Trifolium repens with Digitaria decumbens at one site, and either L. bainesii, T. repens, Macroptilium atropurpureum, Vigna luteola or Stylosanthes guyanensis with Chloris gayana at the other site. (15NH4)2SO4 equivalent to c. 0.3 kg nitrogen ha-1 was added every 4 weeks in an attempt to label the soil mineral nitrogen, and the partitioning of 15N uptake between species used as an estimate of the partitioning of uptake of soil nitrogen. Although two of the legumes (L. bainesii and T. repens) obtained 70–100% of the total 15N uptake at some of the spring harvests, when growth of the associated tropical grasses was limited by low temperatures, none of the legumes obtained more than c. 25% of the annual 15N uptake over a 2-year period. Proportional 15N uptake by the legumes was curvilinearly related to the proportion of legume dry matter yield in the mixtures. Annual 15N uptake by the grass-legume mixtures was up to 360% as high as by the grass control. This had only a small effect on estimates of the proportion of legume nitrogen derived from symbiotic fixation, but caused a large discrepancy in isotopic estimates of nitrogen transfer. The proportion of fixed nitrogen in the legumes averaged 94% at one site and 92% at the other, nearly always exceeded 80%, and was independent of legume yield.

Fertilised pastures of 4 varieties of Setaria sphacelata (var. splendida and cvv. Nandi, Narok and Kazungula), pangola grass (Digitaria eriantha ssp. pentzii) and kikuyu grass (Pennisetum clandestinum cv. Whittet) were grazed rotationally over a 3-year period at Beerwah, south-eastern Queensland. Each variety was grazed on a 5-weekly rotation by 2 leaders for 7 days; a follower herd including 2 testers grazed the regrowth for 7 days. The numbers of followers were subjectively adjusted weekly, depending on feed available. The experiment was designed to provide adequate quantities of forage of maximum potential quality to the leaders throughout the year. There were large and frequently consistent differences between the grasses in yield and its various components, leafiness, leaf and forage density, rate of development of leaf dry matter, and time and intensity of flowering under the management system imposed. Of particular interest were the high leaf yield of Narok, high overall yield of Narok and var. splendida, low rate of development of leaf area index of pangola grass and kikuyu grass, and high forage density of kikuyu grass. These differences might be expected to have an effect on pasture quality, carrying capacity and optimal management strategy.

Animal production was measured over a six-year period from grass-legume pastures receiving different maintenance rates of phosphorus and potassium fertilizers, and continuously grazed at stocking rates of 1.23, 1.65 and 2.47 beasts ha-1. The pastures consisted of the grasses Chloris gayana, Digitaria decumbens, Paspalum dilatatum and P. commersonii and the legumes Desmodium intortum, D. uncinatum, L otononis bainesii, Macrop tilium lath yroides and Trifolium repens. Annual applications of 250 kg ha-1 superphosphate (9.6 per cent P) significantly increased animal production above applications of 125 kg ha-1 superphosphate, but there were no significant effects of increase in level of applied potassium (as KCI) beyond 63 kg ha-1. The greatest liveweight gain per hectare was obtained at a stocking rate of 2.47 beasts ha-1 but rate of liveweight change was more variable than at the other stocking rates. Conversely, daily liveweight gains per head and rate of turnoff were greatest at 1.23 beasts ha-1. Both stocking rates and levels of superphosphate affected quality of carcase produced, and there was a significant positive correlation of liveweight gain and legume content of the pastures.

A complex pasture mixture was grazed by cattle at three stocking rates (1.23, 1.65 and 2.47 beasts a hectare) at Beerwah, south-eastern Queensland over six years. There were four fertilizer treatments involving two rates of superphosphate and three of KCl. On the basis of species frequencies, the seven major soil types on the area fell into two groups, the podzolic soils in one and the humic gleys in another. The podzolics were favoured by Chloris gayana, Desmodium intortum, D. uncinatum and Lotononis bainesii, whereas Paspalum dilatatum and Trifolium repens were more frequent on the humic gleys. The light stocking rate favoured Paspalum commersonii, Digitaria decumbens and both Desmodium species. At the high stocking rate P. dilatatum, T. repens and L. bainesii were more prominent, as were weeds in general. Application of KCl had no effect on botanical composition but the K content of soils and plants increased with increasing levels of application. As the superphosphate rate was increased, the proportion of all sown species except P. dilatatum and L. bainesii increased and the weed content decreased. Over time, the legume content in all plots decreased, weeds increased and the grass content remained about the same. The effects were more marked at low rates of superphosphate and high stocking rate. There was a lower legume content and a higher weed content in winter than in summer, but soils and treatments had similar effects in both seasons. Compared with virgin soils, those under fertilized pastures for eight years contained more organic C, available P, K and Ca, total N, and Zn and Fe.

Animal production from pangola grass (Digitaria decumbens) fertilized with 400 and 800 lb nitrogen an acre a year was measured over a two-year period 1965-1967. The nitrogen was applied as calcium ammonium nitrate (20.5 per cent N) in split applications. These were 27.5 per cent of the total in September, December, and February, and 17.5 per cent in April. The experiment was grazed with yearling Hereford heifers on a 'put and take' system at a mean stocking rate of three beasts an acre. A mean annual production of 1139 and 1215 lb an acre was obtained from 400 and 800 lb N applications respectively. Although there was 20-30 per cent greater dry matter production containing approximately 3.5 per cent more crude protein from the higher nitrogen application for most periods of the year, this had little effect on total animal production. The highest rates of liveweight gain were recorded in the period September-December, coincident with the highest crude protein and phosphate levels in plant material. Liveweight losses occurred in winter but were reduced by 50 per cent under the higher rate of nitrogen application. The results of this experiment clearly demonstrate that high levels of animal production can be obtained from pangola grass in a subtropical environment with adequate nitrogen fertilizer, even though marked seasonal differences in growth rate of the grass occur.