The kinetics of starch digestion in 20 sweetpotato samples obtained from Papua New Guinea were investigated using an in vitro procedure based on glucometry. Irrespective of the cultivars (3-mun, Carot kau-kau, Wahgi besta, Nillgai, Baiyer kaukau, and 1-mun), provinces, farmers and locations, the samples exhibited biphasic digestograms, possibly due to initial impediments to amylolysis and/or transport of digestion products by non-starch components (e.g. plant cell walls). The biphasic digestograms were segmented into two monophasic digestograms using a second-order polynomial (average r(2) = 0.723; p < 0.001), and a modified first-order kinetic model adequately described the digestograms (average r(2) = 0.984, p < 0.001; mean relative deviation modulus, MRDM, =8). Generally, the digestion rate constants (initial, K-i = 3.2-5.2 x 10(-3) min(-1); final, K-f = 3.8-18.4 x 10(-3) min(-1)) indicated faster digestion during the final segment after the initial impediments to digestion had possibly been overcome. K-i was significantly (p < 0.05) different among the samples, which did not significantly (p < 0.05) differ in K-f. A logistic model, which treated each digestogram as a single process, also adequately described the biphasic behaviours (average r(2) = 0.994, p < 0.001: mean relative deviation modulus, MRDM, =15). The digestion parameters from the segment and logistic approaches significantly (p < 0.05) differed among the sweetpotato samples. The rate of digestion from the logistic model was significantly (p < 0.05) related to the initial and final rates of digestion from the segment approach (K = 3.99 K-i + 0.65 K-f: r(2) = 0.456; p < 0.001). Irrespective of the approach, maximum starch digestion was less than 100% in some samples to indicate the presence of resistant starch type 1,2 or both. The present study is the first to model biphasic starch digestograms. Crown Copyright (C) 2010 Published by Elsevier Ltd. All rights reserved.