Narli Plain in eastern Turkey is a 26,500-ha irrigation area brought into a single irrigation scheme between 1975 and 1978. Soils are clay loam and clays. Field observations have shown that compaction limits root development in some areas. To evaluate this, 15 soil profiles were sampled in fallow land after wheat in a rotation that involved cotton, wheat, sugar beet, and maize. Soil structure was measured using bulk density, porosity, and penetrometer resistance. Structure was described using soil thin sections with image analysis and scanning electron microscopy. Soil profiles were ranked on quality. This identified three structural groups; well structured, moderately poorly structured, and very poorly structured.Well structured soils at all depths had densities that were usually less than 1.25 Mg m−3and were always less than 1.4 Mg m−3, which equated with macroporosity greater than 18% and 10%, respectively. Poorly structured soils had densities always greater than 1.44 Mg m−3, and commonly greater than 1.5 Mg m−3, equating with a macroporosity of 9% and 5%, respectively. These values, together with data on pore sizes and morphology, correspond with acceptable versus limiting porosity. An intergrade between these limits described moderately poorly structured soils. The degree of compaction was best explained by considering its spatial distribution. This showed that the most severe compaction occurred within 2 km of the river, near the principal town, and between an archaeological site and the river. This land has been used for 3000 to 4000 years. In contrast, well structured soils occurred in the southwest away from the main settlement and on land reclaimed from marshland in 1958 and cultivated for about 40 years. Land not belonging to these groups has been cultivated for about 400 years and is located further from the river and the settlement than the severely compacted soils and has a moderately poor structure. Poor correlations were found between soil structure and clay mineralogy and calcium carbonate. Clay content was significant but accounted for only 21% of the variation. The role of organic matter was clearer despite small but statistically significant numerical differences between moderately poorly structured soils together with very poorly structured soils (0–40 cm), compared with well structured soils.The more poorly structured soils had a mean organic matter content of 1.45% compared with 2.1% for well structured soils (0 to 40 cm). Thin sections suggested that the soil structure in compact soils had been produced by tillage, partial or total aggregate collapse, and from shrink-swell activity rather than nonplant biota. This is interpreted by the presence of vughs, planar voids, and curvo-planar voids that partially isolate former peds. Biopores and microaggregates comprising coprogenic fragments are rare. A sequence of steps for the formation of severe compaction is suggested with possible remedial actions identified. This work demonstrates that an holistic approach to soil management is required to remediate compaction and that this should take into account the historical aspects of agricultural development and good practice in machinery management.