Abstract
Total dietary fiber (TDF) was measured in large lima, roman, black turtle, light red kidney, white navy, pinto, black-eyed, and soya beans and in chick peas by the Mongeau rapid method (A), the Prosky method (B), and the Lee method (C). When the samples were soaked and cooked according to package instructions (gentle boiling, 95°C), TDF values by method A were all within 19.7-22.1%, except for black-eyed beans (9.9%) and chick peas (11.3%) (g/100 g, cooked dry matter). For large lima beans (20.0-21.3%) and soya beans (19.2-19.7%), TDF values by methods A, B, and C were in agreement. For 7 samples, however, TDF values were up to 81% higher by method B (17.4-34.7%) and up to 122% higher by method C (21.1-39.8%) than those by method A (P ≤ 0.01). For 6 legumes, TDF values by method C were 15-28% higher (P ≤ 0.013) than by method B. White navy beans were analyzed also after different cooking conditions, varying from no cooking to autoclaving for 15 min at 120°C. TDF values by method A were independent from cooking conditions and remained between 20.2 and 22.4%. For navy beans cooked at 95°C, TDF values by method B (up to 34.7 ± 1.4%) and C (up to 39.8 ± 0.3%) were unpredictable, but autoclaving at 120°C reduced them to about 22%. Incorporation of a pan-creatic amylase in methods B and C consistently decreased the aforementioned analytical discrepancies, as did the absence of cooking. Only autoclaving (for at least 15 min at 120°C) fully restored agreement among methods A-C.
