Attempts to improve the accuracy of experimentation in the Griggs-type piston-cylinder deformation apparatus are most welcome and the recent study by Kirby and Kronenberg <1984> on the deformation of clinopyroxenite has the potential for useful quantitative manipulation of rheological data. However, there are many apparent inconsistencies and data-handling inaccuracies in this work that require clarification before further use can be made of the data. The following comments relate specifically to (1) the presentation and/or selection of data, (2) rheological relationships, (3) semibrittle behavior, reproducibility and instability, and (4) single versus multiple mechanisms of deformation.