The Cotton Program has evaluated the Zellweger Uster HVI short fiber measurement over the past four classing seasons. Prior to the 2000 classing season, a revision was made to the short fiber algorithm by Zellweger Uster in an attempt to improve measurement repeatability. Results of the revised algorithm were mixed, with some increase in overall within-lab reproducibility and some decrease in between-lab reproducibility. Overall within-lab reproducibility (tolerance of 1.0) was 61% for year 2000 crop samples compared with 58% for 1999 crop samples. Between-lab reproducibility (tolerance of 1.0) dropped to 50% for 2000 from 52% for 1999.

The predicted short fiber measurement, based on HVI length and uniformity index, has been under evaluation since 1998 (Knowlton, 1999). The predicted short fiber evaluation for year 2000 crop samples resulted in within-laboratory reproducibility of 76% (tolerance of 1.0), up from about 75% for the 1998 and 1999 evaluations. Between-laboratory reproducibility for year 2000 crop samples averaged 74%.

An acceptable short fiber content measurement must have greater accuracy than currently available in either of the above methods in order to provide information regarding short fiber normality for a given HVI length and uniformity index. Until this level of short fiber measurement accuracy is achieved, there is no new information provided beyond what is already known in the HVI length and uniformity index measurements. In the current classification system, where two-specimen tests are utilized, the conclusion after four years of evaluation is that the predicted short fiber measurement provides the best estimation of short fiber currently available.