For a number of reasons, at the outset of the research, the influence of coalescence on the data was ignored. It was initially felt that it was not necessary to include coalescence explicitly in the model. Of course, the model for growth does allow coalesced data to be used, after transformation.
However, it was realised, subsequent to the analysis of the 200kN data set in particular, that coalescence could be a significant factor. The main factor of concern was that the growth only model did not allow for the possibility of coalescence. Thus inferences from the model may be incorrect for the situation where coalescence is a major factor.
In addition to the concerns above, it had been recognised by this author in other research [29] that failure mechanisms are sometimes due to general structural breakdown, caused by large families of cracks throughout a specimen.