In compression, on the other hand, cell wall buckling is most probably the dominating mechanism of deformation, at least for the lower density foams. For higher density foams, the cell walls are both smaller and thicker, suppressing buckling. This explains why the compression strength is that much lower than the tensile strength for the lower density foam. The normalised fatigue data in compression and shear are shown in Fig. 8b and c. They do not collapse into one single generic curve. In compression there is no correlation at all with the two lower densities having a different slope than WF200. However, the WF110 and WF200 are pretty close in terms of normalised stress (Fig. 8b) but with different slopes. In shear we see that the WF110 and WF200 normalised fatigue life curves have the same slope, but at slightly different magnitudes. The discrepancy could either be some influence from cell wall buckling or simply that the density scaling is different in shear.An open pending question that cannot be answered at this point is how the failure mechanism in shear fatigue loading actually looks and eventually leads to failure for the lower density foams. The face/core interface separates (debonding failure) after a certain number of loading cycles, but why at the interface and what is the local failure mechanism at cell size level? This is a question for future research.