The 104 reference cottons for maturity described last year have been sampled (10 samples per cotton). The samples have been tested on both the AFIS and the AFIS PRO (2 replications of 5,000 fibers per sample). This totaled 100,000 fibers per reference cotton per instrument. The two AFIS instruments are in quite good agreement. The shapes of the length distributions are close. Nevertheless, the length distributions with the AFIS PRO tend to show more short fibers and consequently less long fibers.

The next question is related to the possibility of calibrating externally the AFIS with our reference cottons. We have not tried to calibrate the distributions themselves because of a still unsolved statistical problem: “how can you calibrate a distribution?”. Until we can find an answer to this question we could at least calibrate the averages values.