The structural voids in vertisols contain easily available water for plants and their volume can be calculated from the shrinkage curve. Access by plants to that water depends also on the geometric arrangement of the pores so that the water can flow through them. We have devised a method for studying the structural porosity by casting the pores in resin. The intraprism pore space of wet soil clods is impregnated with a UV fluorescent polyester resin under vacuum. When this has set we use the swelling properties of the clay to separate the clay matrix from the resin. A cast so obtained is the real three‐dimensional solid reproduction of the structural porosity. This representation of the pore system is easier to study than results from computerized reconstitution of the three‐dimensional space from two‐dimensional images of soil in thin sections. Channels, packing pores and planar voids can be observed directly in three dimensions as the method saves the integrity and continuity of pores as small as 10 μm in diameter. The geometry of the cast shapes agrees with the interpretation of shrinkage and moisture characteristic curves. The method offers direct qualitative observation of pore organization and volume measurements of the intraprism structural porosity in vertisols.

In many swelling clay soils several types of porosity can be defined: matric, structural and crack porosities. Measuring the hydraulic conductivity associated with the matric porosity is of major interest because soil movements are governed by matric water flows. Our purpose was to determine the matric conductivity of a vertisol from Guadeloupe (French West Indies) under laboratory and field conditions. We used an Eulerian flow description to measure the conductivity of natural vertisol clods in the laboratory with a drying method. In order to take into account soil movements in every direction, we introduced a tensorial analysis of soil deformations and hydraulic conductivity. The instantaneous profile method was used in the field under wetting conditions. Using soil layer thickness transducers we described water flows in a material coordinate system that can deal with volumetric soil deformations. For wet soil the ratio between the hydraulic conductivity measured in the field and that measured at the laboratory was around 10. However, because the spatial variation of the water content was large, no final conclusion could be drawn to explain this discrepancy. We showed that the gradient of the overburden potential was smaller than that of the matric potential in the field experiment.