The formation of adcumulates necessitates the continued growth of primocrysts down to low porosities. Gravitationally driven viscous compaction at the base of a crystal mushy layer on the magma chamber floor, driven by the weight of the mushy layer itself, is commonly suggested as a significant process acting to drive out interstitial liquid and promote adcumulate formation. Compaction necessitates viscous deformation, by either dislocation creep or diffusion-controlled processes such as pressure-solution: many studies suggest that the foliations preserved in cumulates are a consequence of recrystallization during compaction, completely overprinting primary magmatic fabrics. We test the compaction hypothesis by looking for microstructural evidence of viscous deformation. A detailed examination of cumulates from the Skaergaard intrusion, East Greenland, demonstrates only limited crystal plastic deformation, with no correlation between the extent of dislocation creep and the calculated volume fraction of trapped liquid left in the cumulates. Although the evidence for diffusion-controlled deformation is often cryptic, there is an anti-correlation between apparent aspect ratio of plagioclase and the extent of adcumulate crystallization, contradicting previous hypotheses involving transposition of original magmatic fabrics by dissolution–reprecipitation. This is supported by the spatial distribution of compositional zoning in plagioclase, which demonstrates that pressure-solution or related diffusion-controlled processes were insufficient to obscure primary magmatic fabrics. The Skaergaard adcumulates did not form by viscous compaction. Instead we suggest that they formed by primary processes involving mass transport in a thin mushy layer. Compaction is most likely to occur in slowly cooled intrusions in which the bulk magma crystallizes abundant dense minerals. We present preliminary observations of microstructures in norites from the lower Main Zone of the Bushveld Intrusion, South Africa, and in plagioclase-rich cumulates from the Fe–Ti oxide-rich Baima Intrusion, SW China. The evidence for dislocation creep in both intrusions is unambiguous, although deformation was insuffi- cient to obliterate all traces of the primary magmatic fabrics and unlikely to have been sufficient to significantly reduce the volume of interstitial liquid.