ABSTRACT
The impact of a smaller sample of whole meal wheat (5.5 or 6.0 g) to replace the official 7.0 g used in the Falling Number test was investigated using samples of Canada Western Amber Durum (CWAD) and Canada Western Red Spring (CWRS) wheat. Use of either of the smaller sample sizes resulted in a significant shortening of the analysis time of the test. Reproducibility studies, using high and low falling number (FN) CWRS and CWAD samples, with three analyses per day over six days indicated no appreciable change in the coefficient of variation for the test using a smaller sample size. The maximum daily standard deviation of three replicates, 25.7 sec, was observed using the official 7.0-g moisture-corrected CWAD sample and was not significantly different from the 23.0 sec value obtained for a 5.5-g sample of CWRS. The maximum average standard deviations observed over the six days of analysis were 13.7, 7.7, and 5.3 sec for the 5.5-, 6.0-, and 7.0-g sample sizes, respectively, and were all associated with the sound, high FN, CWRS sample. While the use of 5.5 g allowed significant differentiation between high and low FN CWRS samples, the ability to discriminate high and low FN CWAD samples was lost at this sample size. FN analysis of CWRS (n = 144) and CWAD (n = 141) at 7.0 vs. 6.0 g, yielded correlation coefficients of 0.95 and 0.88, respectively. Regression analysis indicated a ±14.1 sec error associated with estimating a 7.0-g FN value using 6.0 g of CWRS which increased to 26.4 sec for CWAD. Particle size analyses of whole meal after grinding indicated that the harder CWAD wheat fractured into a significantly greater percentage of larger particles than the corresponding CWRS. This difference may be a contributing factor to the greater variance associated with the CWAD 6.0-g test and the inability to differentiate sound and slightly sprouted wheat at 5.5 g. Use of the 6.0-g method, stopping the test after 200 sec, would also be applicable for screening purposes at commercial facilities where normal testing times (350–450 sec) could be reduced to meet designated quality requirements. This would offer a savings of 2.5 min/sample over the conventional method.
