The ~SN methods are potentially accurate for measuring N 2 fixation in plants. The only problem with those methods is, how to ensure that the ~SN/14N ratio in the plant accurately reflects the integrated tSN/14N ratio (R) in soil which is variable in time and with soil depth. However, the consequences of using an inappropriate reference plant vary with the level of N 2 fixation and the conditions under which the study was made. For example, the errors introduced into the values of N 2 fixation are higher at low levels of fixation, and decrease with increasing rates of fixation. At very high N 2 fixation rates, the errors are often insignificant. Also, the magnitude of error is proportional to the rate of decline of the 15N/14N ratio with time. Since N 2 fixation in most plants would be expected to below 60%, the question of how to select a good reference plant is still pertinent. In this paper, we have discussed some of the criteria to adopt in selecting reference plants, e.g. how to ensure that the reference plant is not fixing N2, is absorbing most of its N from the same zone as the fixing plant, and in the same pattern with time, etc. In addition, we have discussed ~SN labelling materials and methods that are likely to minimize any errors even when the fixing and reference plants don't match well in certain important criteria. The use of slow release ~SN fertilizer or 15N labelled plant materials results in slow changes in the ~SN/~4N ratio of soil, and is strongly recommended. Where L~N inorganic fertilizers are used, the application of the fertilizer in small splits at various intervals is recommended over a one-time application. The problem with the reference crop, which has sometimes discouraged potential users of the 15N methods, is surmountable, as discussed in this paper. * Figures in parenthesis represent relative ranking made using each method.