Nodulation, N2 fixation (estimated by 15N natural abundance methods) and dry matter production were studied in a lucerne (Medicago sativa) crop managed for hay production at Ginninderra Experiment Station, A.C .T. Measurements were taken in the year of establishment and during two subsequent growing seasons. There were three treatments: (1) no inoculation and no annual fertilizer applied, (2) initial inoculation and superphosphate applied annually, (3) no inoculation, superphosphate applied annually and ammonium sulfate periodically. Before planting and after each growth season, soil was analysed for extractable mineral nitrogen, total nitrogen and the 15N natural abundance of this nitrogen, to the depth explored by lucerne roots. Before planting, no appropriate root-nodule bacteria (Rhizobium meliloti) were detected in the soil and initially plants were nodulated only in the inoculated treatment. Thereafter nodulation increased on the other treatments. Eight months after sowing there were no differences between treatments in numbers of R. meliloti g-l soil or in nodulation. In the third growing season, almost 30 kg ha-1 (dry wt) of nodules were recovered to a depth of 25 cm. These nodules were primarily located on fine, ephemeral roots and many appeared to be renewed after cutting of the lucerne. In the year of establishment, dry matter yields (0% moisture) totalled 3 to 4 t ha-1 in three hay cuts. In succeeding years, total yields were in the range 10 to 13 t ha-1 in four or five cuts per season. Nitrogen removed in the harvested lucerne reached 340 to 410 kg N ha-lyr-l in the second and third years and between 65 and 96% of this N arose from N2 fixation, depending on the method of calculation used. Poorer dry matter production and N2 fixation in treatment 1 in the third growing season was attributed to an insufficient supply of available phosphorus. Fixed N removed in Lucerne hay from treatment 2 totalled at least 640 kg N ha-1 in the three years of the experiment. Also, there were substantial increases in soil nitrogen due to lucerne growth. Although soil compaction made the quantification difficult, at the end of the experiment it was estimated that there was at least an extra 800 kg N ha-1 in the total soil nitrogen under lucerne compared to strips of Phalaris aquatica grown between the lucerne plots. It was concluded that lucerne contributed at least the same amount of fixed nitrogen to the soil as was being removed in the harvested hay.
