Quantitative information on the fate and efficiency of nitrogen (N) fertiliser applied to coarse textured highly calcareous soils in semi-arid farming systems is scarce but, as systems intensify, is essential to support sustainable agronomic management decisions A glasshouse study was undertaken to trace the fate of N fertiliser applied to wheat (Triticum aestivum L. cv Yitpi) grown on a reconstructed profile (0 to 600 mm) of a grey highly calcareous (> 35% CaCO3) sandy loam soil. Two watering treatments were applied (drier and wetter) equivalent to low (decile 2, 179 mm) and medium (decile 5, 234 mm) growing season rainfall for a location with typical semi-arid environment in southern Australia. 15 N-labelled urea fertiliser (35.4 mg N/pot) was applied in a split application - at sowing and 70 days after sowing, followed by immediate watering or watering after delay of 1 week. Recovery of N fertiliser in grain (30 to 52%) was greater for the wetter treatment, and when water was applied immediately following fertiliser application. It was also similar for N applied at sowing and N applied during crop growth. Overall, more than 40% of the urea fertiliser N remained in the soil at anthesis, largely in the top 100 mm, indicating little movement of fertiliser N down the profile even with application of water. Losses of urea fertiliser N (13-24%) were considered relatively small given the highly calcareous nature of the soil; and were significantly greater from N applied during growth compared to at sowing, particularly where watering did not immediately follow application. There was no effect of fertiliser N on grain yield due to sufficient available N in soil at sowing (139mgN/pot), but N concentration and DM of stubbles was increased. Watering treatment did not affect shoot dry matter production up to anthesis, although root weight was higher in the wetter treatment, and grain yield was 9% greater. It is concluded that the potential for N losses from urea applied to highly calcareous coarse textured soils in semi-arid agricultural areas appears relatively low. Further, where there are relatively large amounts of plant available N present at sowing, a strategy of delayed or withheld applications of N to manage economic risk may have minimal effects on grain production in seasons with drier than average rainfall.