In 1911, Gebhardt used a photoelastic model to relate mechanical stresses to the ossification pattern of the chondroepiphysis. Pauwels later conducted a photoelastic study using the same model geometry to develop a theory that the secondary ossific nucleus originates at a position of high‐magnitude hydrostatic pressure where the shear stresses are zero. We conducted two‐dimensional finite element analyses of the model used by Gebhardt and Pauwels. We demonstrate that Pauwels's photoelastic results are correct but are based on the imposition of incorrect boundary conditions. When more realistic boundary conditions were used, the finite element results changed dramatically. These results suggest that (a) the ossific nucleus appears in an area of high shear (deviatoric) stresses;​ (b) the edge of the advancing ossification front (zone of Ranvier or ossification grove) also experiences high shear stresses;​ and (c) the joint surface, where articular cartilage forms, is exposed to high‐magnitude hydrostatic compression. These findings support the theory proposed by Carter and associates that intermittently applied shear stresses (or strain energy) promote endochondral ossification and that intermittently applied hydrostatic compression inhibits or prevents cartilage degeneration and ossification.