Accurate estimation of crop evaporation from a range of soil types is fundamental to the continued improvement of irrigation management. In this experiment soybean crop evaporation was measured using two weighing lysimeters, one with an undisturbed block of Hanwood loam (L1), the other with undisturbed Mundiwa clay loam (L2). Although both soils have clay profiles the Hanwood loam was much more freely draining. A fresh water-table (EC =0.01 S m-1) was maintained 1 m below the soil surface of each lysimeter after 18 days from sowing (DFS). The crop (vat. Chaffey) was sown on 18 November 1985 both in and around the lysimeters and was harvested 136 days later. Early crop growth was slower in L2 but growth stages after 60 DFS were similar to L1. Estimated leaf area indices exceeded 3 about 51 and 56 DFS for L1 and L2 respectively and were both greater than 9 at 90 DFS. The crop was well watered until at least 110 DFS. However daily rates of evaporation (E) from LI noticeably exceeded those of L2 beginning around 50 DFS and increased to 30% greater for the period 70 to 1 | 5 DFS. This difference was not due to incorrect calibration. Plant water status measurements were generally similar although covered leaf water potential and foliage temperature values indicated that plants on L1 were less well hydrated than on L2. The causes of the E difference are not known, but it was observed that plants in L2 were about 0.1 m shorter than the surrounding plants which were similar in stature to those in L1. It was speculated that this difference in height created a shelter effect which reduced the net radiant energy absorbed by the canopy at low sun angles and reduced the wind speed controlling the turbulent exchange of water vapour. This study highlights the lack of understanding that exists about the effect that small discon-