ABSTRACT
If the part of a cube or other sample of a paste or mortar that has been attacked is only a layer extending inwards a short way from the surface, there is not much hope of obtaining significant information by grinding up the entire sample and examining it by some standard procedure, such as XRD or thermal analysis. This nevertheless appears to have been done in at least one classic and otherwise well planned investigation. Similarly, while expansion measurements on samples of a single size may be useful as a comparative procedure, they cannot provide information of a fundamental nature.
