Rheological measurements made by a number of laboratories on the test fluids M1, A1 and S1 have revealed excellent agreement in shear flow (both steady state and oscillation). Extensional flow data, however, exhibited a wide divergence between different techniques, although the data from methods such as the spin line and methods such as opposed jet fell within distinct regions. The question now arises as to why this should be the case. The essence of the problem lies in the fact that the measurements being made in all the different techniques are transient. At no time does the extensional viscosity approach an equilibrium value. Consequently, the time scale of the deformation process has a profound effect on the value of this transient extensional viscosity. The postulate that the extensional viscosity results trace out a path along a surface characteristic of any given fluid would seem to offer an explanation for these discrepancies. Three-dimensional plots of transient extensional viscosity η e, Hencky strain ε, and time t show that surfaces can be generated over which move the results from different experiments at different times. It is even possible to accommodate results using averaging techniques, even though they appear to move backward along the time axis. Results which are consistent on the three-dimensional surface plot may appear totally unrelated projected onto a two-dimensional η e vs. ε (or ε) graph. For the first time results from one fluid, S1, are shown to lie on just such a surface. In addition, it was even possible to generate a surface for the polymer melt itself (polyisobutylene), using an Instron tensile tester, and then to compare the results with those from a commercial controlled strain rate extensional rheometer. Excellent agreement was found between the two sets of data.