Field experiments (20 m 2 plots) were conducted to compare Azolla and urea as N sources for rice (Oryza sativa L.) in both the wet and dry seasons. Parallel microplot (1 m 2) experiments were • -1 ° conducted using 15N. A total of approximately 60 kg N ha was applied as urea, Azolla, or urea plus 1 Azolla. Urea or Azolla applied with equal applications of 30 kg N ha-at transplanting (T) and at maximum tillering (MT) were equally effective for increasing rice grain yields in both seasons. Urea at 30 kg N ha-i at T and Azolla 30 kg N ha-1 at MT was also equally effective• Urea applied by the locally recommended best split (40 kg at T and 20 kg at MT) gave a higher yield in the wet season, but an equal yield in the dry season. The average yield increase was 23% in the wet season, and 95% in the dry season. The proportion of the N taken up by the rice plants which was derived from urea (%NdfU) or Azolla (%NdfAz) was essentially identical for the treatments receiving the same N split. Recovery of 15N in the grain plus straw was also very similar. Positive yield responses to residual N were observed in the succeeding rice crop following both the wet and dry seasons, but the increases were not always statistically significant. Recovery of residual lSN ranged from 5.5 to 8.9% for both crops in succeeding seasons. Residual recovery from the urea applications was significantly higher than from Azolla in the crop succeeding the dry season crop• Azolla was equally effective as urea as an N source for rice production on a per kg N basis.

The soybean is a legume which has an ability to supply its major nitrogen need by the biological nitrogen fixation (BNF) process. This process is made possible by nodules formed in their roots, colonized by Rhizobium sp.bacteria. An accurate estimation of N gained by BNF is necessary to predict the increase or decrease of chemical fertilizer-N requirements to increase soybean production. Among several methods, the 15 N method was used to estimate the ability of legumes to perform BNF. The study involved soybean var. Willis (W) and a completely non-BNF soybean var. CV, which is termed as a standard crop. The standard crop is nonnodulated soybean, but it has the same main physiological traits with var. Willis. The aim of this study was to determine whether 15 N-labelled fertilizer with different %a.e. given to nodulated and non-nodulated soybean would not be of significant consequences for the calculation of N-BNF of W. The treatments applied were different rates of urea (20 kg N/ha and 100 kg N/ha) combined with different atom excess percentages (%a.e.) 15 N (2% and 10%). Thus, the combination of treatments were as follows:(1) W-ll (20 kg N; 2% a.e); (2) CV-hl (100 kg N; 2% a.e); (3) W-lh (20 kg N; 10% a.e); (4) CV-hh (100 kg N; 10% a.e); (5) CV-ll (20 kg N; 2% a.e); (6) W-hl (100 kg N; 2% a.e); (7) CV-lh (20 kg N; 10% a.e); (8) W-hh (100 kg N; 10% a.e). The result of the experiment showed that a high %a.e. with a low rate of 15 N and a low %a.e. with a high rate of N should be used to study the %N-BNF of nodulated plants.