Presented here is a serious, and by no means fault searching, re-examination of what has been known as "calibration" in cotton fiber testing. The calibration, being an essential pre-requisite for the measurement process, is in fact an inverse problem involving a process of identifying the performance of an instrument, arising from the direct causal relationship it represents. Effect of using an assembly which is non-constant and random, like a clamp of cotton, as the calibration standard on the possible magnitude of measurement errors is made explicit to recommend planned further studies on the subject in the hope of reducing the uncertainty of the cotton quality measurements.