We determined that the number of insect fragments, quantified using the standard flotation method, in flour milled from wheat infested with larvae, pupae, or preemergent adults of the lesser grain borer, Rhyzopertha dominica (F.), was proportional to infestation level. Wheat infested with a single preemergent adult contributed 28 and 10× as many fragments as wheat infested with a single larva or pupa, respectively. Using regression models that were developed from these data, we predicted that the maximum infestation level that would result in flour with fragment counts below the Food and Drug Administration defect action level (75 fragments/50 g of flour) was 0.95 and 1.5% (380–640 infested kernels/kg of wheat) for pupae and larvae, but it decreased to 0.05% (20 infested kernels/kg) when the grain was infested with preemergent adults. We also reexamined the accuracy and sensitivity of near-infrared spectroscopy (NIRS) for detecting insect fragments in flour by testing three different NIR spectrometers. NIRS-predicted numbers of insect fragments were correlated with the actual number of fragments. NIRS is less precise than the standard flotation method, but it is rapid, nondestructive, does not require extensive sample preparation, and could easily be automated for a more sophisticated sampling protocol for flour based on prescreening samples with NIRS followed up by use of the standard flotation method when necessary.