Predictions from laboratory tests of the compression behaviour of peat from 14 sites are compared with full-scale field loading at five sites. Data presented confirm the heterogeneous nature of the deposits. However, for typical engineering works, calculations based on laboratory test data are likely to give reasonable predictions of the magnitude of immediate and primary compression. Standard (20 mm thickness) samples may give misleading data on time for primary consolidation. Thicker samples (e.g. 50 mm) should be used. Sampling by conventional samplers, as used for mineral soils, can cause densification of the peat, resulting in underestimation of actual settlement. Block samples or sample tubes with serrated cutting edges are recommended for peat soils. It was found that the data presented follow the Cα/Cc law of compressibility. There is also is some evidence to suggest that the H2 scaling law may be applicable. Good correlations were found between vertical yield stress (pvy′) and compression index (Cc) and index parameters such as water content (wi). Conventional staged construction with surcharge loading may be successfully applied to peat soils as long as adequate drainage exists to permit consolidation over reasonable time intervals.