(A) The percent reduction in yield strength associated with adding resorption cavities to cancellous bone, compared to the case without cavities, depended on bone volume fraction (p<0.0001, n=40) and (B) SMI (p<0.0001) when cavities were targeted to regions of highly-strained tissue (filled symbols) but not when they were distributed randomly (unfilled symbols, p>0.43). (C) When resorption cavities were distributed randomly, the percent reduction in yield strength associated with adding such cavities depended on mean trabecular thickness (Tb.Th, p<0.0001).