Effects of falling number sample weight on prediction of alpha-amylase activity

Reports vary on the effects of falling number (FN) sample weight on test pr ecision, reproducibility, and predictability of alpha -amylase activity. St raight grade flours of 200 samples (25 cultivars x 2 locations x 2 N, level s x 2 repetitions) were assayed for alpha -amylase activity and FN. Locatio n significantly affected alpha -amylase activity and FN values. The coeffic ients of variation (CV) for the FN tests were 5.75, 2.12, 1.93, 1.72, 4.27, and 14.47%, when assayed with sample weights of 7, 6, 5.5, 5, 4.5, and 4 g , respectively. The FN test with the greatest reproducibility between sampl e replicates (lowest LSD and highest ratio of range/LSD) was also produced using the 5-g sample weight. By reducing FN sample weight from 7 to 5 g, FN values that averaged 350 sec, considered essentially sound, averaged 215 s ec, thus shortening the FN test time by an average of 2 min and 15 sec when assaying sound wheat flour. The results suggest a review of the 7-g stipul ation of AACC Approved Method 56-81B for FN in favor of reduced sample weig ht.