The famed Carrara marble has been used as outside-exposed, thin, facade panels that have failed because of out-of-plane distortions; a phenomenon referred to as dishing, cupping, and bowing. Theories have been advanced to explain that phenomena. This paper expands the explanations, and is based upon the relationships of thermal hysteresis the stone is known to possess and stone dimensions.

The marble consists of a mosaic of fine, essentially equi-axial calcite crystals devoid of defined crystal boundaries. Calcite has a known anisotropic thermal response wherein the crystal expands (as would be expected) in the direction of the ‘c’ crystallographic axis, but contracts (as would not be expected) in the direction of the three ‘a’ crystallographic axes, which are oriented normal to the ‘c’ axis. Very slight repositioning of the crystals result because of dislocations along crystal boundaries, which interfere with the return of the crystals to their original locations. The resulting slight increase in volume of the marble is accumulative with continued thermal cycling.

When the front of panels undergo more thermal cycles and/or reach higher temperatures than the back of panels, the result is a differential volume increase front to back with the front expanding more than the back. Based upon laboratory experimental work, when the ratio of lateral dimensions of the panels to their thickness is appropriate, out-of-plane distortions reconfigure the panels in the form of a permanent slight dish.

The physical effects of the thermal hysteresis on the marble causes; dishing, increased volume, increased porosity, decreased strengths because of stress along crystal boundaries and, ultimately, disintegration of the marble into sugar-like granules.

Marbles having coarser grain sizes and mosaic textures do not undergo similar thermal hysteresis effects as the Carrara marble because there are fewer crystals to interact.