An attempt was made to re-assess the results of irrigation experiments carried out in the UK during the last 30–40 years in order to develop functional relationships between crop yield and actual wateruse. It was first necessary to develop a model, based on relatively simple functions, which could be used to predict actual water-use (evaporation and transpiration). The derivation of this model, known as WATYIELD, is reported here, whilst its application and validation for sugarbeet and for potatoes are described in subsequent papers. Simple functions relate components of evaporation to the proportion of the incoming solar radiation intercepted by the crop canopy, reference crop evapotranspiration, the precipitation and irrigation amounts and the soil water deficit. Transpiration is also related to soil hydraulic characteristics, rooting density and potential evaporation rate.

As a prerequisite for developing crop-yield/water-use functions for sugarbeet using the results of historical irrigation experiments, it was necessary to develop a water-use model which could operate with a limited data set. The general form of this model has been reported by Wright et al. (1994). In this paper the development and validation of the model for the sugarbeet crop is described.

The canopy was modelled in terms of the leaf area index and the relative interception of incoming solar radiation using functions based on thermal time and time. Four phases of growth were identified: emergence, slow-growth, fast-growth and full-canopy. An empirical drought factor was included to allow for the effects of water stress on canopy development during the slow- and fastgrowth expansion phases. Root development was described using a three phase model: initial (temperature dependent), linear and maximum depth (both time dependent).

Independent data previously reported from Broom's Barn Experimental Station were then used to validate the model in terms of its capacity to predict crop canopy development, with and without drought stress, soil water extraction at different depths and soil water deficits during the season. The study confirmed the validity of the model for predicting the water-use of sugarbeet.

The growth of kernels at selected positions along the spike of barley was studied using three models: segmented, logistic and Gompertz. The segmented model divides the growth period into three segments: an initial constant stage, a middle period of ‘linear growth’, and a final constant stage. The identification of a period during which growth is approximately linear - the ‘linear phase’ - was estimated from the curvilinear functions by imposing some criterion to determine the period of ‘linear growth’ in order that the results were similar to those from the segmented model. From this a model of final growth, growth rate and duration of growth as a function of kernel position was developed and fitted to data from four different thinning treatments at five sowing dates. The model of final growth and growth rate was shown to have a family of gamma functions. The analysis of these models showed that there were marked interactions between sowing and thinning treatments for growth rate, less so for grain yield, but there was no substantial interaction for duration.

Experiments were conducted in glasshouses at IRRI, Philippines, during 1987 and 1988 to identify varietal differences in apical development using six rice cultivars having extreme characters; Jirasar 280 (early senescence), Neo-Dunghan 1 (late senescence), Macunting (small seeds), ABB (large seeds), C12474–1 (long culm), and B581–A6–458 (short culm); and five cultivars, IR20, IR22, IR34, IR54, IR64, of the same growth duration.

The length of time from physiological panicle initiation (PI) to flowering varied from 25 to 50 days. The onset of physiological PI occurred when the fourth penultimate leaf was fully exserted. The visual panicle initiation stage (VPI) coincided with spikelet differentiation in all cultivars. Nitrogen top-dressing at 55–57 days before the anticipated maturity date (the present N management recommendation) did not coincide with the actual panicle initiation stage in any of the cultivars and the major yield determinants (i.e. the primary and secondary branches) therefore differentiated in conditions of internal nitrogen shortage. Spikelet abortion took place both before and after flowering. Reduction of pre-flowering abortion is important in increasing rice grain yields and could be achieved by improved N management. It is important to recognise cultivar differences and periods of loss of potential yield in developing efficient N management protocols.

Leaf development was studied in eight related grasses, grown in field swards cut at 5-week intervals, during the year of sowing and the subsequent year (1989 and 1990). The rate of leaf expansion was in the order Westerwolds ryegrass > Italian ryegrass (Lolium multiflorum), Italian ryegrass × meadow fescue > hybrid ryegrass > perennial ryegrass × meadow fescue, meadow fescue (Festuca pratensis), tall fescue (Festuca arundinacea) and perennial ryegrass (Lolium perenne). The order of grasses was similar, but not identical, for rate of leaf appearance, rate of leaf extension, weight of leaf blade emerging per shoot per week and rate of increase in length of exposed leaf sheath, and the order was approximately the reverse for weight per unit area of emerging leaf blade. The area per leaf blade increased greatly between May and October of the year of sowing, particularly in Westerwolds, Italian and hybrid ryegrasses and Italian ryegrass × meadow fescue. Area per leaf blade in tall fescue increased greatly between May and July of the year of sowing and May–July of the subsequent year. Rate of leaf expansion in meadow fescue was much higher in May of the year after sowing than in the previous May.

The effects of particle size on the nitrogen degradability of four tropical forages were studied in 1991 in Guadeloupe. Samples of pangola grass (Digitaria decumbens) 22 and 47 days old, Gliricidia sepium and Leucaena leucocephala were prepared as follows: grasses were (1) freshly cut with scissors to 0·5–1·0 cm in length and frozen at – 18 °C; or for later study were dried at 80 °C and ground to pass a (2) 0·5 mm, (3) 10 mm or (4) 20 mm screen. Nitrogen degradability (ND) was determined by placing samples in nylon bags with two different pore sizes (25 and 46 μm) which were then put into the rumen of cows for 2, 4, 8, 16, 24 and 48 h. The kinetics of nitrogen degradation were examined using Ørskov's model. Particle losses through the nylon bags, dry matter (DM) and nitrogen solubility of the samples were also measured in vitro. The sample preparation and the type of forage were the main sources of variation in the rapidly degradable nitrogen fraction, the slowly degradable nitrogen fraction, the degradation rate, the potentially degradable nitrogen fraction and ND. Nitrogen degradability was 55·8,46·7, 640 and 46·5% for pangola grass (at 22 and 47 days regrowth), Gliricidia and Leucaena samples, respectively. Mean ND was 47·9, 59·4, 56·1 and 49·6% for freshly cut and 0·5, 1·0 and 2·0 mm dried ground samples, respectively. Sample preparation had little effect on nitrogen solubility. For samples dried and ground at 0·5 and 1·0 mm, particle losses were 18·8 and 15·0% of DM, respectively. The insoluble but degradable fraction was 60·8, 51·9, 42·5 and 42·7% for freshly cut and 0·5, 1·0 and 2·0 mm dried ground samples, respectively. The freshly cut material appeared to be suitable for the estimation of ND in tropical forages.

A programme of forage legume evaluation was developed at the International Livestock Centre for Africa's (ILCA) subhumid research site in Nigeria from 1985 onwards. Evaluation included monitoring herbage and seed yields, incidence of diseases/pests, regeneration ability and persistence for three growing seasons. This was followed by an in situ bioassay to determine the direct effect of the sown legumes on maize (Zea mays L.) production in the absence of added nitrogen.

The results for selected accessions from introductions received in 1987 and 1988 are presented. Stylosanthes guianensis CIAT184 (ILCA164) and ILCA15557 and Aeschynomene histrix ILCA 12463 were promising both in terms of biomass production (up to 9·4, 3·4 and 9·6 t/ha respectively) and quality (all had crude protein values of 11% or more), and in terms of beneficial effects for maize production (all plots gave significantly more grain than those with no legume, reaching over 3 t/ha for maize grown on plots after S. guianensis ILCA15557). The use of such appropriate forage legumes to enhance both livestock and crop production in a sustainable farming system is recommended.

Summary Between 1985 and 1988, 14 field experiments at different sites in the UK assessed the effect of applying P, K, Na and Mg in autumn, early spring, late spring or as a split application in both autumn and late spring on the growth, yield and processing quality of sugarbeet. Supplementary experiments compared Kaynitro and Chilean nitrate of soda. None of the treatments had any effect on plant population densities, and any differences in sugar yield could be explained by the leaching of Na. The effects on processing quality were inconsistent and unlikely to be of agronomic or economic significance. Generally, Kaynitro and Chilean nitrate of soda performed no better than the other fertilizers so it is unlikely that their extra cost would be justified. Apart from Na, the timing of nutrient application appeared to be unimportant, even on soils of low fertility. However, Na, being very mobile, was best applied in spring.

Substrates for mushroom cultivation were prepared, following a 2 day mixing and blending process, in bulk tunnels under a controlled temperature regime using forced ventilation. The temperature regime was based on a conventional bulk tunnel composting process, i.e. pasteurization at 60 °C for 6 h, followed by a conditioning phase at 47 °C until the substrate was clear of ammonia. With the exception of ammonia, which increased with increasing compost nitrogen content, this process did not result in strong odours. The substrates were ready for inoculation with mushroom ‘spawn’ 7–12 days after the initial mixing of the compost ingredients. Increasing the compost nitrogen content from 1·1 to 2·5% of the dry matter by increasing the quantity of deep litter poultry manure added to straw in the ingredients resulted in a greater subsequent yield of mushrooms. Further increases in the substrate nitrogen content resulted in prolonged tunnel processing times, substrate desiccation, incomplete clearance of ammonia from the substrate and subsequently low or no mushroom yields. Substrate bulk density at the time of spawning decreased with increasing nitrogen content, but was increased by chopping the straw ingredient. Mushroom yields from composts prepared with barley straw were significantly lower than those from wheat straw composts, at equivalent nitrogen contents. Supplementation of prepared substrates with the proprietary protein-rich ingredient, Betamyl 1000, increased yields by 13·6%.

A field experiment was conducted in Canterbury, New Zealand to investigate the effect of six leguminous and non-leguminous grain crops on soil N fertility over a 12 month period (March 1989 to March 1990). All crops had an overall negative N balance during their growing season. A greater amount of soil N was removed by barley, rape and lupins (104–119 kg N/ha) than by field beans, field peas or lentils (50–74 kg N/ha).

Net N mineralization was measured in all treatments between residue incorporation and the start of winter. With the exception of the lupins, accumulation of mineral N in the soil profile before the start of winter drainage was greater following leguminous (mean 124 kg N/ha) than non-leguminous crops (mean 80 kg N/ha).

Cumulative apparent leaching losses over the autumn/winter were largely a reflection of the mineral N content of the profile before the start of drainage. Excluding lupins, leaching losses declined in the order fallow > legumes > non-legumes (110 > 72 > 37 kg N/ha respectively). The anomalous results for the lupins were attributed to the incorporation of a large amount of woody residues after harvest which may well have resulted in extensive net N mineralization occurring later in the autumn.

Over a 12 month period, all treatments showed a decline in N fertility (110–160 kg N/ha), although compared with barley, the total loss of soil N was 10–40 kg N/ha less following leguminous crops.

Growth of the following spring wheat test crop was affected by the preceding crop. Grain yield, grain N yield and total N yield were significantly related to the mineral N content of the soil at the end of leaching, and to a measure of net N mineralization during the growing season of the test crop.

Artificially reared sambar (tropical) deer and red (temperate) deer were confined indoors in metabolism cages and fed chaffed lucerne hay ad libitum for 4-week periods during summer and winter at Flock House Agricultural Centre, New Zealand, during 1992. Measurements were made of voluntary feed intake (VFI), apparent digestibility, faeces particle size distribution, eating and ruminating time and the rate of chewing during eating and ruminating. Red deer reduced VFI (kg DMI/day) markedly from summer to winter, associated with a reduction in the duration of each eating bout. Sambar deer slightly increased VFI over this time, associated with an increase in chewing frequency. Digestive efficiency was similar in both species, and the critical particle size for leaving the rumen was passage through a 1 mm sieve for both species. Time spent eating/g DMI was greater for sambar deer than for red deer during summer, but there was no difference during winter. Relative to red deer, sambar deer consistently spent more time ruminating/g DMI, and spent a greater proportion of total ruminating time as daytime ruminating and had more daytime ruminating bouts. Duration of each ruminating bout (min) was similar for the two deer species, but sambar deer had less chews/bolus ruminated but more rumination boli/h than red deer. Differences between sambar deer and red deer were more pronounced in ruminating than in eating behaviour, and sambar deer may have evolved a different rumination pattern to break down low-quality tropical forages more effectively.

Lotus pedunculatus was grown under high fertility conditions and its nutritive value was determined in a feeding trial with sheep at Palmerston North, New Zealand in 1989. The condensed tannins (CT) accounted for 5·5 % of lotus dry matter (DM) and its effect on digestion was evaluated by giving an intraruminal infusion of polyethylene glycol (PEG) to six of the sheep (PEG group). PEG preferentially binds with CT so that the lotus becomes essentially CT-free.

The experiment was carried out with 14 sheep (six PEG and eight ‘Tannin’) held in metabolism crates indoors and given freshly cut lotus hourly, for 32 days. This paper presents data relating to carbohydrate and mineral digestion, together with aspects of rumen function.

Digestibility of lotus DM was 68%, and the digestibility of fibre was not affected by CT. Infusion of PEG increased rumen concentrations of NH3 and volatile fatty acids (P < 0·001) but effects on molar ratios of VFA were inconsistent with time. CT reduced rumen degradation and absorption of sulphur and increased net absorption of both phosphorus and zinc, but other effects on mineral digestion were small.

Although the lotus was offered at c. 90% of ad libitum, intakes of the tannin sheep began to decline after c. 15 days of feeding and were c. 12% lower than those of the PEG sheep at the end of the trial (P < 0·05). At slaughter, rumen pool sizes were similar for the two treatments but the Tannin sheep had a lower fractional outflow rate, which suggests a slower rate of digestion in the rumen. Growth rate and wool production were similar for sheep on both treatments. It is concluded that the CT in Lotus pedunculatus grown under high fertility conditions had little effect on fibre and mineral digestion but the depression in DM intake reduced its nutritive value for sheep.

Fourteen young wether sheep were fed freshly cut Lotus pedunculatus as a sole diet to examine the effects of condensed tannins (CT; 55 g/kg lotus DM) on nitrogenous aspects of digestion. The experiment was carried out indoors at Palmerston North, New Zealand over 32 days with one group of sheep receiving an intraruminal infusion of polyethylene glycol (PEG; 100 g/day) to preferentially bind CT (PEG group) so that the lotus was essentially ‘CT-free'. The other sheep, not given PEG, were termed the ‘Tannin’ group.

The principal effects of CT were to increase the flow of feed nitrogen (N) to the abomasum despite a 12% reduction in DM intake of the Tannin sheep. Rumen microbial N turnover rate was slower in Tannin animals than in those receiving PEG (1·86 v. 2·63/day) but microbial N flux to the abomasum was similar in both treatments. The proportion of N intake disappearing from the rumen was lower in Tannin (0·13) than in PEG sheep (0·26) and the N digestibility was 0·67 and 0·81 for the respective treatments (P < 0·001).

The beneficial effects of CT in reducing rumen degradation of feed protein were negated in part by a reduction in fractional absorption of amino acids (AA) from the small intestine. Fractional absorption of essential AA was 0·66 in Tannin and 0·79 in PEG sheep; values for non-essential AA were 0'59 in Tannin and 0·73 in PEG groups. Amino acid concentrations in blood were similar for both groups, but Tannin sheep had lower plasma urea concentrations, a more rapid plasma urea turnover rate and a higher irreversible loss than those receiving PEG. Growth hormone concentrations in plasma were similar for both treatments.

The effects of sucrose supplementation on nitrogen kinetics and energy metabolism were examined in sheep fed lucerne hay cubes using a 15N isotope dilution method and balance and respiration trials in Tsukuba, Japan in 1988. Sheep were fed lucerne hay cubes (1183 g DM/day), with or without 204 g/day sucrose, at 2 h intervals from continuous feeders. Supplementation with sucrose decreased urinary N excretion (P < 0·01), resulting in an increase in N retention from – 1·1 to + 2·0 g N/day (P < 0·01). Supplementation with sucrose resulted in no change in faeces and methane energy, a decrease in urinary energy (P < 0·01) and an increase in heat production and energy balance (P < 0·01). Sucrose supplementation also resulted in lower rumen ammonia (P < 0·05) and plasma urea concentrations (P < 0·05) and reduced urinary urea excretion (P < 0·01). The fermentation of sucrose in the rumen resulted in a decrease in rumen pH (P < 0·01) and in the acetate: propionate ratio (P < 0·05). Sucrose supplementation increased the proportion of urea transferred to the rumen (P < 0·05), non-ammonia N (NAN) concentration in the rumen (P < 0·001) and NAN flow from the rumen to the lower digestive tract (P < 0·001). Urinary allantoin excretion rate increased with sucrose supplementation (P < 0·05). The plasma glucose concentration was unchanged but plasma insulin concentration was increased with sucrose supplementation (P < 0·05). The influence of energy-rich supplements, such as sucrose, on N kinetics and the mechanism of the increase in N retention with sucrose supplementation are discussed.

Two populations of perennial ryegrass (PRG) (Lolium perenne L.) were selected for low (LS) and high leaf shear breaking load (HS) in 1984 at DSIR, Palmerston North. Leaf shear breaking load for the LS group was c. 41% lower than for the HS group and the LS selection had significantly shorter, narrower leaves and smaller leaf cross-sectional areas than the HS selection. LS leaves were c. 27% less resistant to shear than HS leaves per unit of cross-sectional area. The lower leaf shear strength in the LS selection was associated with a significantly lower concentration of sclerenchyma tissues in the leaf cross-sectional area compared with the HS selection. The total shear load required to break down the same dry weight of leaves to 1 mm particle size (leaf index of masticatory load) was calculated and shown to be influenced by the differences in leaf morphology. LS leaves were estimated to have a 21% lower index of masticatory load than HS leaves when both were grown under optimum climatic conditions.

A series of experiments was conducted at Palmerston North, New Zealand, during 1988–91 to compare the efficiency of chewing during eating and rumination, rumen fractional outflow rate (FOR), voluntary organic matter intake, liveweight gain and wool production in sheep fed either low (LS) or high (HS) leaf shear breaking load perennial ryegrass (PRG). The LS ryegrass had a 13% lower mean leaf shear breaking load and ingestion rates tended to be higher than for HS PRG, but no consistent significant treatment differences were observed in rate of particle breakdown, rumen FOR, voluntary intake or animal performance. It was concluded that selection for reduced leaf shear breaking load per se did not improve feeding value. The total shear load required to reduce a unit dry weight of PRG leaf to < 1 mm particles (index of masticatory load; IML) differed by only 3% between LS and HS PRG in this study, due to higher leaf length:dry weight ratios for LS PRG. It is therefore suggested that IML, which takes into account both leaf shear breaking load and associated changes in leaf morphology may be a better criterion for selection than leaf shear breaking load alone in breeding programmes to improve the feeding value of perennial ryegrass.

This group, which is concerned with the applications of mathematics to agricultural science, is sponsored by the Biotechnology and Biological Sciences Research Council. It was formed in 1970, and has since met at approximately yearly intervals in London for one-day meetings. The twenty-fifth meeting of the group, chaired by Dr T. M. Addiscott of the Institute of Arable Crops Research, Rothamsted Experimental Station, was held in the Wellcome Meeting Room at the Royal Society, 6 Carlton House Terrace, London on Friday, 15 April 1994, when the following papers were read.