Transmittance spectra (500 to 950 nm) and reflectance spectra (550 to 1700 nm) were analyzed to determine if they could be used to distinguish aflatoxin contamination in single whole corn kernels. Spectra were obtained on whole corn kernels exhibiting various levels of bright greenish-yellow fluorescence. Afterwards, each kernel was analyzed for aflatoxin following the USDA-FGIS Aflatest affinity chromatography procedures. Spectra were analyzed using discriminant analysis and partial least squares regression. More than 95% of the kernels were correctly classified as containing either high (>100 ppb) or low (<10 ppb) levels of aflatoxin. Classification accuracy for kernels between 10 and 100 ppb was only about 25%, but these kernels do not usually affect total sample concentrations and are not as important. Results were similar when using either transmittance or reflectance, and when using either discriminant analysis or partial least squares regression. The two-feature discriminant analysis of transmittance data gave the best results. However, for automated high-speed detection and sorting, instrumentation that uses single-feature reflectance spectra may be more practically implemented. This technology should provide the corn industry with a valuable tool for rapidly detecting aflatoxin in corn.

The accuracy of using near‐infrared spectroscopy (NIRS) for predicting 186 grain, milling, flour, dough, and breadmaking quality parameters of 100 hard red winter (HRW) and 98 hard red spring (HRS) wheat and flour samples was evaluated. NIRS shows the potential for predicting protein content, moisture content, and flour color b* values with accuracies suitable for process control (R2 > 0.97). Many other parameters were predicted with accuracies suitable for rough screening including test weight, average single kernel diameter and moisture content, SDS sedimentation volume, color a* values, total gluten content, mixograph, farinograph, and alveograph parameters, loaf volume, specific loaf volume, baking water absorption and mix time, gliadin and glutenin content, flour particle size, and the percentage of dark hard and vitreous kernels. Similar results were seen when analyzing data from either HRW or HRS wheat, and when predicting quality using spectra from either grain or flour. However, many attributes were correlated to protein content and this relationship influenced classification accuracies. When the influence of protein content was removed from the analyses, the only factors that could be predicted by NIRS with R2 > 0.70 were moisture content, test weight, flour color, free lipids, flour particle size, and the percentage of dark hard and vitreous kernels. Thus, NIRS can be used to predict many grain quality and functionality traits, but mainly because of the high correlations of these traits to protein content.

Various whole‐kernel, milling, flour, dough, and breadmaking quality parameters were compared between hard red winter (HRW) and hard red spring (HRS) wheat. From the 50 quality parameters evaluated, values of only nine quality characteristics were found to be similar for both classes. These were test weight, grain moisture content, kernel size, polyphenol oxidase content, average gluten index, insoluble polymeric protein (%), free nonpolar lipids, loaf volume potential, and mixograph tolerance. Some of the quality characteristics that had significantly higher levels in HRS than in HRW wheat samples included grain protein content, grain hardness, most milling and flour quality measurements, most dough physicochemical properties, and most baking characteristics. When HRW and HRS wheat samples were grouped to be within the same wheat protein content range (11.4–15.8%), the average value of many grain and breadmaking quality characteristics were similar for both wheat classes but significant differences still existed. Values that were higher for HRW wheat flour were color b*, free polar lipids content, falling number, and farinograph tolerance. Values that were higher for HRS wheat flour were geometric mean diameter, quantity of insoluble polymeric proteins and gliadins, mixograph mix time, alveograph configuration ratio, dough weight, crumb grain score, and SDS sedimentation volume. This research showed that the grain and flour quality of HRS wheat generally exceeds that of HRW wheat whether or not samples are grouped to include a similar protein content range.

Reflectance and transmittance visible and near‐infrared spectroscopy were used to detect fumonisin in single corn kernels infected with Fusarium verticillioides. Kernels with >100 ppm and <10 ppm could be classed accurately as fumonisin positive or negative, respectively. Classification results were generally better for oriented kernels than for kernels that were randomly placed in the spectrometer viewing area. Generally, models based on reflectance spectra have higher correct classification than models based on transmittance spectra. Statistical analyses indicated that including near‐infrared wavelengths in calibrations improved classifications, and some calibrations were improved by including visible wavelengths. Thus, the color and chemical constituents of the infected kernel contribute to classification models. These results show that this technology can be used to rapidly and nondestructively screen single corn kernels for the presence of fumonisin, and may be adaptable to on‐line detection and sorting.

Cereal Chem. 87(6):511-517Fusarium Head Blight (FHB), or scab, can result in significant crop yield losses and contaminated grain in wheat (Triticum aestivum L.). Growing less susceptible cultivars is one of the most effective methods for managing FHB and for reducing deoxynivalenol (DON) levels in grain, but breeding programs lack a rapid and objective method for identifying the fungi and toxins. It is important to estimate proportions of sound kernels and Fusarium-damaged kernels (FDK) in grain and to estimate DON levels of FDK to objectively assess the resistance of a cultivar. An automated single kernel near-infrared (SKNIR) spectroscopic method for identification of FDK and for estimating DON levels was evaluated. The SKNIR system classified visually sound and FDK with an accuracy of 98.8 and 99.9%, respectively. The sound fraction had no or very little accumulation of DON. The FDK fraction was sorted into frac-* The e-Xtra logo stands for "electronic extra" and indicates that Figures 1 and 3 appear in color online.