A short introductory treatment of the basic theory of the mercury strain-gauge plethysmograph is presented. Some physiological aspects of this type of plethysmography are then discussed which, among other things, illuminate the difficulty in discriminating between skin-flow and muscular flow. A detailed calculation of the elastic influences follows and shows that one cannot generally expect a cancellation of these influences when comparing measurements with the calibration. This part illuminates further the difficulties in discriminating between skin-flow and muscular flow. As a result, such discrimination should be regarded as dubious. Some other sources of error are also discussed briefly. Two proposals for a new design are given. Finally, the Appendix shows the values of the gauge stretch which will give cancellation of the elastic errors in a simple model. The actual values cannot be used in practice but they indicate that the amount of stretch is quite critical.