The studied pore systems are described as array of chambers which are connected by smaller throats. N₂ adsorption is used to calculate the size of the pores, whereby no distinction between throat or chamber type is possible with this method. Assuming mercury entrapping in ink-bottle type pores (pores that are connected to an external surface through smaller pores only) being the dominant process for mercury snap-off during extrusion and applying multi-cycle MIP, the calculation of the size of the entrances of these ink-bottles is possible. It is shown that similar results also may be derived from mercury extrusion data by applying a contact angle correction for the retracting mercury meniscus. A good agreement of the pore size distribution of the connected, non-ink-bottle type pores derived from either N₂ sorption or mercury intrusion is obtained.

Samples of cement paste and mortar are analysed. A significant difference between cement paste and mortar regarding the neck entrances of ink-bottle type pores is found and attributed to the coarse pore space around the aggregates, the interfacial transition zone.