In a study of methods of measuring gains and losses o£ nitrogen in grazed pastures, the following aspects were investigated:
(1) methods of measuring net nitrogen changes in large plots of grazed pastures by direct sampling, (2) the use of 15N as a tracer to estimate legume nitrogen fixation, and (3) the use of 15N to measure losses of nitrogen under field conditions.
(1) In grazed pasture plots 3 to 6 ac in area, there were large differences between pastures in the number of soil or plant samples needs for a chosen level of precision. Using stratified random sampling, the estimated number of 1.8 in. diameter soil cores needed for a least significant difference of 50 lb N/ac in the 0-6 in. soil layer varied from 190 in a setaria pasture on a yellow podzolic soil to 1070 in a green panic pasture on a sedentary clay soil. The number of 5 x 1 lk2 quadrats needed for a least significant difference of 10 lb N/ac in the plant component (cut at ground level) varied from 10 in a Townsville lucerne/spear grass pasture to 154 in a green panic pasture. The number of soil or plant samples needed was not closely related to the nitrogen content of the respective components.
Statistically, the most efficient method of measuring nitrogen changes was to subdivide the plots into strata and repeatedly sample the same set of sampling sites within these strata, taking one sample per site on each occasion. Calculations based on estimates of costs and variance components for different stages of sampling showed that the above method was also the most economical.
The precision of measurements of net nitrogen changes in two continuously grazed pastures was similar to that predicted from the sampling studies.
In one pasture, a fertilized Townsville lucerne/spear grass pasture at Rodd's Bay, Queensland, the soil/plant/animal system lost 55 lh N/ac (S.E. ± 18) during 2½ years in spite of an estimated addition of at least 18 lb N/ac by Townsville lucerne. In the other pasture, a fertilized Siratro/grass pasture at Samford, Queensland, there was no change in the nitrogen content of the soil/plant/animal system (S.E. ± 17 lb N/ac) despite an estimated addition of approximately 45 lb N/ac by Siratro, It was thought that a large part of the compensating losses of nitrogen was due to transfer of nitrogen by cattle to camping areas in the pastures and volatilization of nitrogen from dung and urine.
(2) The 15N technique appeared to work quite well for the separation of uptake of soil nitrogen and symbiotic nitrogen fixation by legumes grown in pots, though the basic assumption that the legume and grass take up soil nitrogen and labelled fertilizer nitrogen in the same ratio has not yet been tested.
The partition of 15N uptake between grass and legume in microplots within large plots (several acres) of pasture was affected by the proportion of legume (on a dry matter basis) in the mixtures. The partition of 15N uptake indicated that in pastures containing up to 20 per cent Townsville lucerne or up to 60 per cent Siratro, the legume obtained less than 20 per cent of the available soil nitrogen.
(3) Studies of the recovery of 15N from microplots in a pasture indicated that reliable measurements of 15N recovery in field experiments can be made for a reasonable cost of 15N and without excavating the entire soil mass from the microplots. There was no evidence that the core sampling method gave biased results. The method of using core sampling to measure recovery in unconfined microplots was discussed in detail.
Under conditions considered unfavourable for leaching or denitrification, 23.5 per cent (S.E. ±3.7) of the applied 15NO3-N was apparently lost from the soil/plant system (0-12 in. soil depth) after ten months.