Spatial data would have undoubtedly led to a different treatment of the research question. That said, the lack of spatial data has focused the research to concentrate on deducing as much as possible from the available information. It is acknowledged that the collection of detailed spatial data constitutes a difficult and tedious task. In principle, however, some spatial information could in the future be recorded at the same time as the lengths. Ideally, such information may take the form of detailed locations of the cracks; an alternative might consist of dividing the specimen into regions, noting in which region each crack resides and recording the microcrack density for that region. This would be useful in determining different rates for different regions, and for estimating reliability based on future coalescence. It would also be a much more reasonable expectation of engineering practitioners than the recording of detailed locations. It is possible that an analysis subsequent to the collection of such data may demonstrate that the crack parameters have also a dependence on local region, which may be incorporated into the hierarchical model.