Abstract Automated flow control coupled to differential GPS guidance systems in aerial topdressing aircraft will allow variable rate (VR) fertiliser strategies to be applied on hill country farms. The effectiveness of these strategies will be enhanced with the use of remotely sensed hyperspectral data to categorise and quantify the farm landscape in greater detail. The economic benefit of a variable rate fertiliser strategy in comparison to a single rate (blanket) strategy was evaluated for a case study Whanganui hill country station. The analysis illustrates the robustness of a VR strategy in the face of volatile returns in that it produced a higher 10 year cumulative net present value (NPV) and remained at a positive advantage at three different stock gross margins, in comparison to a blanket approach. The effectiveness of hyperspectral imagery for defining effective pasture areas to assist development of more precise variable rate fertiliser applications, compared to the current visual classification from farm photography is discussed. Keywords: economic benefit, variable rate fertiliser, hyperspectral data

New Zealand pastoral farming has benefited greatly from the application of phosphorus (P) and sulphur (S) fertilisers supplied in the main by superphosphate (SSP). The long-term fertiliser trial at Winchmore, mid-Canterbury, New Zealand was set up in 1952 and has yielded a wealth of data on the effect of fertiliser, grazing by sheep and flood irrigation on pasture production. The trial was initially (1952 1958) designed to measure the effect of no fertiliser and SSP applied each year at 188, 376 and 564 kg ha(-1). From 1958 to 1980, SSP applications were stopped to the 564 and one half of the 376 kg SSP y(-1). The cessation of fertiliser decreased clover content and increased the proportion of weeds and low-fertility grasses. The decline in production once fertiliser application ceased followed a curvilinear pattern, but never reached the low production of the no fertiliser treatments even after 20 years. In 1980, the residual treatments were changed to compare a reactive phosphate rock (RPR)/S treatment as well as an intermediate (250 kg ha(-1)) rate of superphosphate. The 0, 188 and 376 kg SSP ha(-1) have now been unchanged for 60 years, while the 250 kg SSP or equivalent in RPR/elemental S have remained unchanged for 30 years. The 188 and 250 kg SSP ha(-1) y(-1) treatments have shown that without any nitrogen fertiliser, ryegrass and clover will persist in irrigated pastures and result in high levels of pasture production (11-12 t ha(-1)) for up to 60 years.

Grassland net primary productivity, carbon (C) allocation to roots, root production and subsequent turnover time in grazed pastures have significant implications for modelling landuse effects on global carbon dynamics. The objective of this work, using the decline of assimilated C-13, was to quantify C allocation to roots, root production and turnover time in response to long-term fertiliser and irrigation treatments. The treatments on the pastures on a long-term research site at Winchmore, Canterbury, New Zealand were 0 or 375 kg ha(-1) y(-1) superphosphate under irrigation and, on a nearby irrigation experiment, unirrigated or irrigated (with fertilisation of 250 kg ha(-1) y(-1) superphosphate) when the soil water content fell to 20% w/w (50% available moisture). The pasture treatments were pulse-labelled using (CO2)-C-13 within portable gas-tight enclosures. Separate micro-plots were (CO2)-C-13 pulse-labelled in late spring, summer and autumn. Below ground net primary production (206 g C m(-2) y(-1)) was similar in unfertilised, unirrigated, fertilised and irrigated pastures, despite marked differences in above ground production. Unfertilised/irrigated and unirrigated/fertilised treatments had greater root biomass, root C allocation and longer root turnover time (1.9 y and 2.0 y, respectively) than fertilised and irrigated treatments (1.3 y). These root turnover times appeared to be consistent with the improved substrate quality attributed to the species found in the irrigated and fertilised treatments.

Total soil concentrations of cobalt (Co) and manganese (Mn) were determined in a range of pastoral soils sampled from four different regions of New Zealand. Concentrations of Co and Mn extracted from soils with 0.02 M Na(2)H(2)EDTA (pH 7.0) and with 0.05 M CaCl2 were also determined, and, in the case of Co, were determined using both air-dried and field-moist samples. Relationships between total and extractable soil Co and Mn concentrations, and other soil properties, were examined by means of statistical correlation and regression analysis. There were significant correlations between total and EDTA-extractable Co and Mn in both top- and subsoils, indicating a strong geochemical association between Co and Mn in soils during their development. As assessed by CaCl2 extraction, there were only small proportions of soil Co and Mn present as soluble and exchangeable forms in New Zealand soils. It was observed that soluble Co and Mn concentrations could be reasonably well predicted from soil pH and the respective EDTA-extractable metal concentrations. Air-drying of soils increased CaCl2-extractable Co concentrations by an average of five-fold, indicating the likely sensitivity of soluble soil Co concentrations to soil water potential.

Field-trial data from a database comprising records of 804 potassium (K) fertiliser trials were used to define the production functions relating exchangeable soil K (quick test K (QTK) 0-75 mm) to the relative response to fertiliser K applications, for the major soil groups in New Zealand. For all soil groups for which there were sufficient data, the production functions were generally flat in the range QTK 5-10, and thus the estimated relative pasture production at QTK 5 and QTK 10 were similar. The critical QTK levels to achieve 97% maximum production were relatively well defined, being 6 (5-8) for sedimentary soils (brown and pallic) and brown soils, and 7 (5-10) for pumice soils. The data for the allophanic soils were unstable and the best estimate was 6 (5-10). For the remaining soils groups (podzols and raw soils, organic, recent and gley soils) for which there was much less data, the relationships were essentially flat over the range QTK 2-10. The probability of pasture responses to applied K increased as soil QTK decreased from 10. For the sedimentary and volcanic soils (including both allophanic and pumice) the probability was about 70-80% at soil QTK < 2. The comparable probabilities were 50-60% for the recent and gley soils, and 30-43% for the podzols and raw soils. A feature of the response functions was that some trials were not responsive to fertiliser K despite having low soil QTK. In most cases this could not be attributed to soil K reserves as measured by the soil TBK test (sodium tetra-phenol-boron extractable which measures exchangeable K plus plant-available but non-exchangeable K). Other possible reasons for this feature in the data are discussed, including uptake of K from below the soil sampling depth and the temporal effects of clover responses to applied K. Soil K buffer capacities-the amount of fertiliser K over and above maintenance required to increase soil QTK by 1 unit (Delta K)-ranged from 50 to > 150 kg K ha(-1) (average 124) for sedimentary soils. For some soils (developed organic soils, gleyed soils and podzols), fertiliser K had very little effect on QTK (0-75 mm). It is not clear whether these differences are due to differences in leaching of K from the sampling depth, differences between soils in their ability to absorb and retain applied K or indeed the result of errors in the measurement of this parameter. Estimated maintenance K requirements (i.e. the amount of applied K required to maintain soil QTK levels) increased with increasing soil QTK from 4 to 10, from 0-150 kg K ha(-1) yr(-1) to 100-300 kg K ha(-1) yr(-1) in situations where losses of K were extreme due to the removal of all harvested clippings. Given the uncertainties in predicting K responses and the amount of fertiliser K required to correct K deficiency, practical suggestions are offered as to how best to diagnose and manage soil K deficiency. Areas for future research to improve the prediction of pasture responses to fertiliser K are also included.

This study examined the influence of irrigation on soil phosphorus (P) distribution and availability under permanent pasture in New Zealand. Soil samples (0–0.075, 0.075–0.15, 0.15–0.25 m) were taken from a long-term field experiment, which included a dryland and 2 irrigation treatments (irrigated at 10% and 20% soil moisture) that had received 25 kg P/ha annually as superphosphate for 52 years. Corresponding data for soil from an adjacent ‘wilderness’ site that had not been used for agriculture for 54 years were included for comparison. Analyses included total P, organic P, and inorganic P; isotopic exchange kinetics (IEK) was used to determine soil inorganic P pools of differing plant availability. Concentrations of total and inorganic P were greater in soil taken from the dryland treatment than the irrigated treatments at all depths. This was attributed to a combination of decreased pasture growth and P transfer in drainage and off-farm produce. Concentrations of organic P were greater in the irrigated treatments (e.g. 0–0.075 m: 672–709 mg P/kg) than in the dryland treatment (e.g. 0–0.075 m: 574 mg P/kg) as a consequence of increased pasture production and soil biological activity. Inorganic P availability (Cp and E1min) was also greater in the dryland treatment than the irrigated treatments. Furthermore, concentrations of inorganic P in the recalcitrant IEK pool (E>3m = E3m–1y + E>1y) in the 0–0.075 m soil from the dryland treatment (479 mg P/kg) were significantly greater than the 10% irrigated (346 mg P/kg) and 20% irrigated (159 mg P/kg) treatments, which was mainly attributed to physico-chemical reactions that decreased the exchangeability of accumulated inorganic P with time. Despite increased P retention capacity at depth (R/r1, 0.15–0.25 m: dryland 6.6, 10% irrigated 10.2, 20% irrigated 12.8), concentrations of total inorganic P in the 0.15–0.25 m soil layer were lower under irrigation (195–266 mg P/kg) than dryland (354 mg P/kg), which indicated that long-term flood irrigation increased P transfer by leaching. The findings of this study revealed that while irrigation improved the utilisation of applied fertiliser P it also resulted in increased P movement to depth in the soil profile.

A database was constructed comprising records from 2255 pasture phosphorus (P), potassium (K) and sulphur (S) field trials, of which 1799 included one or several rates of P. Subsets of this data were selected based on predetermined criteria to define the relationships between relative pasture production and available soil P (0-75 mm, Olsen P in mu g P cm(-3) soil)-the P production functions-for the major soil groups in New Zealand. These relationships, and their 95% confidence intervals, were defined using Bayesian statistics. For those soil groups for which there was sufficient data, the production functions were well defined and gave reasonably precise estimates of the relative pasture yield for a given Olsen P. For example, for the volcanic soils, the relative pasture production is most likely (P < 0.05) to be in the range 88-94% at Olsen P 25 and 98-100% at Olsen P 50. The shape of the production functions was similar for all soil groups-the relative pasture production increased with increasing Olsen P up to an asymptote-except the pumice soils and the podzols, which showed irregularities. The production function for the podzols was also flatter. There was good agreement between the empirically derived production functions and those generated from a dynamic P model. The Olsen P level required to achieve 97% maximum production was estimated for all soil groups. These ranged from 10 to 45 depending on soil group. The critical Olsen P levels were related to the soil anion storage capacity (ASC, a laboratory measure of P buffer capacity) and to soil volume weight (g cm(-3) of sieved and dried soil), although not strongly. The field measured P buffer capacity (Delta P-F)-the amount of soluble fertiliser P (kg P ha(-1)) required above maintenance to increase the Olsen P (0-75 mm) level by I unit-was estimated for selected trials. There was reasonable agreement between these estimates and those derived from the P model (Delta P-M), and these results indicated that Delta P decreases with increasing Olsen P. The results imply that factors other than those related to soil chemical properties affect the relationship between soil P and pasture production. The factors which determine the relationship between pasture production and soil P are defined and discussed. These were assigned to two categories: those factors which affect the ability of the soil to supply P for plant uptake and those that affect the ability of the plant to acquire soil P. It is concluded that further progress towards improving our ability to predict pasture responses to fertiliser P will depend on quantifying the latter effects. Based on these results and the development of a dynamic P model, an econometric P model was developed for New Zealand pastures which enables consultants to quantify the likely agronomic, financial and investment effects of any given fertiliser strategy on a given farm or block within a farm. This was not previously possible but is essential for the sustainable use of P fertilisers in pastoral farming.

A glasshouse pot trial was conducted using 18 New Zealand grassland soils to assess the effects of soil manganese (Mn) and cobalt (Co) status, Co fertiliser and soil moisture conditions, on the plant availability of soil Co. The uptake of native and applied Co by ryegrass was measured, and attempts made to determine the main factors influencing plant Co concentrations. There were highly significant relationships between ryegrass Co concentrations and total soil Mn, and EDTA-extractable and CaCl2-extractable soil Mn; ryegrass Co concentrations decreasing in a curvilinear fashion with increasing soil Mn levels. Similar relationships were observed between ryegrass Co concentrations and Mn determined in individual soil fractions as determined using a sequential fractionation technique. It was therefore clearly demonstrated that soil Mn status plays a crucial role for soil Co availability. Soils with high Mn contents have a high probability of strong fixation of soil Co and showed negligible responses to Co fertiliser treatment. For the soils in this study, which had a large variation in Mn status, soil Co extractants were poor predictors of Co availability to plants. However, soil Co fractionation data suggests that Co in the organic-bound fraction is probably one of the important sources of supplying Co in soils for plant uptake.

A reactive phosphate rock (RPR) submodel has been incorporated into an integrated decision support package, Overseer(TM), which is a tool to derive site-specific phosphorus (P), potassium (K), and sulfur (S) nutrient recommendations accounting for the key biophysical, economic, and environmental factors that influence fertilizer requirements and effects. The software uses input information readily available to farmers and their consultants and is primarily targeted to fertilizer company representatives and farm consultants.

A modified sodium tetraphenyl-boron method (Na-TPB), which uses an incubation period and boiling with a copper solution to destroy the tetraphenyl-boron-K precipitate (TBK) instead of shaking and resolubilisation with acetone, is proposed as a replacement for Jackson's TBK method. The method as recommended is I g soil in 3 ml of 1.7 M NaCl/0.01 M EDTA solution + 0.2 g Na-TPB and incubated at 20degreesC for between I and 168 h. Testing of the method showed it to be a more rapid, accurate, and cost-effective means of routinely testing soil K reserves than Jackson's test. Correlation of TBK values against Jackson's method were good (r(2) = 0. 88) for short extraction times (I 4 h) but decreased with increasing extraction time. Regressions of TBK values with K-c values, the current method for differentiating K supplying ability between soils, were only fair (r(2) < 0.75). This was due to mechanistic differences between dissolution (K-c-acid extraction) and solubility based (TBK-precipitation) methods.Potassium release curves were established for a range of 24 New Zealand soils for 1-672 h. The bulk of K release (c. 80%) occurred for most soils within the first 48 h with release rates in significant decline after 168 h. Release curves were modelled using a range of simple kinetic models: empirical Elovich , power, exponential, simplified-Elovich, and parabolic diffusion equations were all examined. The quality of fit for each model tended to decrease in that order, although this varied between soil classes. The linearised functions for a number of soils suggested that K release was biphasic and their description could be better approximated by two line segments, corresponding to initial rapid K release with a second phase of more slowly-available K. This release pattern equates approximately to K held predominantly in "wedge zones" and clay mineral interlayers, respectively, and suggests that the kinetics of reserve-K release often involves more than a single mechanism and thus, more than one associated bond energy.Implications for New Zealand reserve-K soil testing methodology are that three measures of K status are required to characterise K release and potential pasture response; exchangeable K (initial), readily-available K (short-term incubation; 1-16 h) and long-term available K (168 h incubation). Currently, insufficient field trial data is available to fully optimise soil test conditions for the readily-available K measure.

The amount of soil carbon is the net effect of carbon inputs and decomposition and is influenced by many factors including the level of herbage utilisation, root production, litter quality, soil fertility and soil moisture status. Long term experiments at Winchmore, in mid- Canterbury, show soil carbon levels increased after initial development, but after 50 years fertiliser input has had no net effect on soil carbon, while increasing the frequency of irrigation has actually resulted in lower soil carbon levels. Pasture development at Ballantrae , in the southern Hawke's Bay, has been associated with a decline in soil C levels, with the high fertility treatment having only slightly higher soil C than the treatment receiving a lower rate of fertiliser. At Tara Hills, in South Island semi-arid high country, increased stocking rate is associated with a decrease in soil C. Contrary to common perceptions increases in pasture and farm productivity do not necessarily result in increased soil carbon storage. Keywords: carbon, fer tiliser, grazing management, irrigation, soil organic matter

Changes in reserve-K status were measured on a number of historical and current pasture trials using a modified sodium tetraphenyl-boron incubation method (TBK incubation). Trials included four P, K, and S trials from North Otago, a long-term (>40 years) P fertiliser trial from Canterbury (Winchmore), a long-term (>20 years) nutrient depletion trial from Manawatu, and a 2-year N x K trial from Southland. While only two trials showed a clear DM response to K, testing by both the original (Jackson's) test and the TBK-incubation method showed changes in soil K status for all soils. The long-term fertiliser trial showed that K reserves declined with normal annual superphosphate additions when no K fertilisers were applied. Failure of net changes in K reserve status to adequately account for K removal in nutrient budgets appeared to be best explained by significant plant uptake of K below the soil sampling zone (0-75 mm).Prediction of a dry matter K response when exchangeable-K levels are near or below critical levels i.e., at Quicktest K (QTK) values < 6-7, is governed by the buffering capacity of soil K reserve. This was linked in this study to measures of both readily available (short incubation; 1-20 h) and long-term available (168-h) K. Soils with a large long-term K reserve can still exhibit a K response under high K demand, but this occurred in only one trial when QTK values were 3 or less, and when short-term available K was also low. The biphasic nature of K release underlies the complications of predicting K response in many sedimentary soils. The main means of classifying and predicting K response in New Zealand has historically been from Metson's soil class K-c values. This study showed that K reserve status for a significant number of New Zealand soils is not adequately represented by these values, and that better prediction of K response could be achieved using the new TBK-incubation method after optimisation against field trial data. Implications for using this method to advance routine soil testing of K are discussed.

Abstract Data on the physical and financial performance of farms are collected annually by the Meat and Wool Economic Service of New Zealand. We used the available data from 1968/69 to 1995/ 96 to calculate a nutrient balance for the model high‐country farm. Our nutrient budget takes account of the loss of nutrients in the form of livestock products (meat and wool), and losses or movement due to leaching and rainfall run‐off. Inputs of nutrients occur with the application of fertiliser, in rainfall, and by legumes fixing nitrogen. While there have been large fluctuations in both the quantity of nutrients applied as fertiliser and the quantity of nutrients removed, the whole‐farm nutrient balance has remained in credit each year for the whole of the time series for all major nutrients. Separate budgets for developed and undeveloped grassland showed that the nutrient balances were in credit for four of the five nutrients in each case; in developed grassland mean potassium balances were slightly negative while in the undeveloped grassland phosphorus balances were slightly negative. The annual surplus of calcium is significantly higher than for other nutrients. The results indicate a need for re‐interpretation of nutrient losses calculated for post‐European settlement as they affect understanding of the effects of grazing on the sustain‐ability of high‐country grasslands. Over the 30 years studied there appears to have been a cessation and probably some redress of nutrient losses from the system. This indicates that losses were greater earlier in the 150‐year period of European settlement than are indicated by the long‐term averages.

Many New Zealand soils are known to produce pastures deficient in cobalt (Co) for grazing ruminants; however, the development of an effective soil diagnostic test for Co is hindered by the lack of knowledge of the forms and distribution of Co in New Zealand soils. A sequential fractionation technique was used to determine the forms and concentrations of native and/or fertiliser derived soil Co in 18 New Zealand grassland topsoils, and the distribution and interrelationships of Co between fractions compared with those for manganese (Mn). Previous studies in other countries have indicated a strong association between Co and Mn in soils. For both Co and Mn, very small proportions were present in soluble and exchangeable forms, and the highest proportions of Co and Mn (mean 45.8% and 35.8%, respectively) were found in the residual fraction. However, there was little similarity overall in the distribution of Co and Mn between fractions. Multiple regression models, and the presence of substantial amounts of Co and Mn in iron (Fe) oxide fractions, indicate that the Fe content of the soil is much more important than Mn in influencing the distribution of Co in the soils. Exchangeable Co and Mn also appear highly sensitive to soil pH. Soil organic matter did not appear to have a major effect on the distribution of Co in soils, apart from Pumice Soils that have low concentrations of Fe and Mn.

The effects of soil properties on the availability of cobalt (Co) for pasture uptake have been studied using a wide range of New Zealand soils. There is a strong positive correlation between total or EDTA-extractable Co and manganese (Mn) concentration. However glasshouse studies have shown that Co availability is inversely related to the Mn concentration. Across a wide range of soil groups, EDTA-extractable Co is a poor predictor of cobalt availability. CaCl2-extractable Co was found to decrease with soil pH and was too pH sensitive to be a good predictor of Co availability. It is concluded that soil Mn plays a crucial role in soil Co status and has significant effects on plant Co uptake because of its involvement in the soil redox system and the scavenging properties of soil Mn oxides. Co deficiency is more likely to occur and Co fertiliser is less effective on soils with high Mn, especially under dry conditions. Keywords: lime, pasture cobalt, soil cobalt, soil manganese, soil moisture, trace elements

Pasture yield responses to phosphorus (P), sulphur (S), and potassium (K) were measured over four to eight years on different yellow-grey earth (pallic) soils at three non-irrigated (Timaru, Kauru, and Claremont soils) and one irrigated (Otiake soil) sites in North Otago. Large pasture yield responses to P at application rates up to 80 kg ha(-1) yr(-1) occurred on three sites with initial soil Olsen P levels of 6-11 mu g ml(-1), but only a small response up to this rate was measured on the Timaru soil with a higher initial soil Olsen P level (16 mu g ml(-1)). The relationship between soil Olsen P and relative annual pasture yield was similar and moderate (r(2) = 0.57) for the average of the three most responsive sites. The highest pasture yield response to S was measured at application rates up to 80 kg ha(-1) yr(-1) on an Otiake soil with an initial soil sulphate S level of 3 mu g g(-1). Smaller pasture yield responses at rates up to 40 kg S ha(-1) yr(-1) occurred on Kauru and Timaru soils with initial soil sulphate S levels of 8 and 4 mu g g(-1), respectively. In four of the eight years there was a pasture yield response up to 10 kg S ha(-1) yr(-1) on the Claremont soil with an initial soil sulphate S level of 9 mu g g(-1) Soil sulphate S only accounted for a small proportion (27%) of the average variation in relative yield. Three of the four sites had medium to high initial soil quick test (QT)K levels (9-19) and the fourth (Otiake soil) had a low level (4). Over four to eight years, the soil QT K in the nil K treatments declined to 2-7, but there was a pasture yield response only in the last four years at the lowest level, on the Otiake soil. Soil potassium tetra-phenyl boron levels were increased by K application and showed a lower decline than for no K. There was a moderate to strong relationship measured between relative yield (% of maximum yield) and %P (r(2) = 0.58) and %S (r(2) = 0.83) in mixed herbage on the Otiake soil but not at the other sites. %P was related to Olsen P (P = 0.52-0.69) on the Claremont and Otiake soils and %S was related to soil sulphate S (r(2) = 0.51) for the mean of all four sites. Overall, the results show that North Otago yellow-grey earth soils can be responsive to P, S, and K, and the size of the pasture yield response can be predicted with a reasonable degree of accuracy by the level of available soil P, S, and K, the amount of soil reserve K, and the P, K, and S concentration in mixed herbage.

The distribution of a solution simulating cow urine in soil was measured within 6 hours of deposition, by applying 2 litres of KBr-pyranine dye solution to the soil surface at a rate of 0.2 litres s(-1) followed by excavation and sub-sampling of the area affected by urine in 10-cm depth increments until no further pyranine dye could be detected under UV fluorescence. Preferential flow of simulated cow urine to below 20 cm soil depth occurred in 9 of the 10 soils examined, tip to 68% of applied solution moved below 20 cm, at an average of 17% over all soils examined. Two soils exhibited a high degree of preferential flow and excluding these from the data reduced the average loss to 11%. Although this preferential flow was significant, much of the urine nutrients moving via preferential flow would still be available for uptake by plant roots at depth, and therefore cannot be considered a loss to pasture. The movement of urine to below 20 or 30 cm soil depth was best predicted by saturated hydraulic conductivity calculated as the mean for the 0-30 cm soil layer. Saturated hydraulic conductivity measurements made for any one 5-cm depth increment, particularly the 0-5 cm soil depth, were poor predictors of urine flow to below 20 or 30 cm depth.

The allocation of carbon (C) to plant roots and conversion to soil organic matter is a major determinant of the size of the terrestrial C pool in pastoral ecosystems. The aim was to quantify C allocation to roots in contrasting pastoral ecosystems. Pastures on long-term research sites in Canterbury, New Zealand were pulse-labelled using 13 CO 2 within portable gas-tight enclosures. Sites included Winchmore (with or without superphosphate fertiliser, and with or without irrigation) and Tara Hills (low, medium or high grazing intensity with continuous or alternating grazing). Separate micro-plots were labelled in late spring, summer and autumn at Winchmore and in spring at Tara Hills. Herbage label 13 C recoveries were greatest one hour after pulse labelling and declined by 21 days, whereas in roots they were initially lower but generally continued to increase until 21 days. The greatest recoveries of 13 C in roots, one hour and 21 days after labelling, were in summer and autumn respectively. The proportion of label 13 C allocated to roots by 21 days was 0.50 in the absence of superphosphate and 0.41 in the superphosphate treatment, and was 0.39, 0.43 and 0.51 respectively in spring, summer and autumn. Irrigation had no significant effect on root allocation. The low stocking rate at Tara Hills, which had the greatest herbage biomass, also had greater total 13 C, tussock herbage 13 C and root 13 C recoveries than the higher stocking rate treatments. Inter-tussock root recovery and allocation of 13 C to roots increased with increasing stocking rate, whereas tussock root allocation was greatest in the high and least in the medium stocking rate treatment. By 21 days there was a greater inter-tussock and tussock root recovery and lower inter-tussock herbage recovery in the continuous than in the alternating grazing management treatment. The root allocation was generally greater in the continuous than in alternating grazed treatments, except for tussocks one hour after labelling where the reverse was the case. In conclusion the 13 C pulse labelling showed pasture plants allocate more C to roots with low soil fertility, high grazing intensity, continuous grazing, and in autumn.

Decision support software for the evaluation of phosphorus (P) and sulphur (S) fertiliser strategies can assist farm consultants and their clients to determine pastoral agriculture fertiliser policies. The underlying dynamic model is based on the P and S cycles in the soil-plant-animal system, including fertiliser inputs and soil- and animalmediated losses. Initial soil fertility is estimated from the Olsen P and phosphate-extractable organic S soil tests, and recent fertiliser history. In the S sub-model organic S and elemental S pools are considered. Animal production responses to fertiliser are estimated from relationships between soil P and S status, fertiliser inputs, pasture relative yield and stocking rate. Maintenance and economically optimum fertiliser strategies can be automatically calculated or the user can enter their own scenarios. Alternative fertiliser policies can be evaluated in terms of nutrient budgets, soil P and S status, pasture and animal production, and economics. Keywords: decision support, economics, fertiliser, model, phosphorus, sulphur

The role of cobalt (Co) and vitamin B 12 in animal nutrition is well understood, but the practical questions of predicting when and where a defnency is likely, and the most effective preventive strategy, remain unanswered in Southland and Otago. The incidence of Co deficiency in the region has recently increased, particularly where the problem had not previously been recognised In a long term maintenance Co topdressing experiment at Woodlands Research Station, pasture Co concentrations have been monitored since 1982 and lamb growth measured in two seasons. Cobalt deficiency has been observed only in the two driest seasons. It is postulated that soil manganese may influence the seasonal availability of Co to pasture plants. The effect of Co topdressing on pasture Co levels in the season of application was very short lived. Rate of cobalt sulphate application affected peak pasture concentration but had little effect on the duration of the pasture response. Lambs grazing treated plots did not gain a substantial reserve of vitamin 8, 2 Timing of Co application in relation to the likely period of deficiency is very Important. Keywords: pasture mineral content, lamb growth rate, vitamin B, 2, cobalt