Relationships between kernel vitreousness and dry matter degradability for diverse corn germplasm I. Development of near-infrared reflectance spectroscopy calibrations

Negative correlations between corn vitreousness and ruminal dry matter and starch degradabilities have been widely reported. To measure corn vitreousness and density more rapidly, Correa et al. [Correa, C.E.S., Shaver, R.D., Pereira, M.N., Lauer, J.G., Kohn, K., 2002. Relationship between corn vitreousness and ruminal in-situ starch degradability. J. Dairy Sci. 85, 3008-3012] initiated the development of near-infrared reflectance spectroscopy (NIRS) calibrations from 47 samples derived from 14 US and five Brazilian commercial hybrids. In this study, we generated more data to add to these NIRS calibrations with the objective of making them more robust. We also evaluated the potential of using Stenvert hardness measurements for NIR calibrations. Thirty-three diverse corn germplasm, sources were grown at University of Wisconsin West Madison Research Station. These included a wide range of endosperm characteristics from opaque 2 (o2) types to densely packed flint types, and a number of intermediates. Harvest was at 1/2 milkline and black-layer maturity stages. Dried kernels from middle portions of ears from 12 selected inbreds, four each from low (0-30%), medium (30-70%), and high (70-100%) vitreousness classifications were used to determine vitreousness by manual dissection and density by water displacement using a pycnometer. Hardness was determined on all 33 inbreds on a 20 g sample using a Stenvert micro hammer-cutter mill with 2 mm screen size and 3600 rpm to measure time to collect ground sample to a set receptacle height (T); total column height (CH); and height ratio of coarse to fine (C/F) particles. The NIRS equations were selected on the basis of high R-2-values (0.90,0.92,0.85, and 0.85) and low SEC (4.85, 0.01, 1.39, and 0.19) and SECV (6.04, 0.02, 1.79, and 0.25), for vitroueness, density, T and CH factors, respectively. Calibrations for vitreousness and density were regarded as the best prediction models compared to stenvert hardness measurements as determined by their RPD values (3.73 and 2.50, respectively). These results show that NIRS can be used as a screening tool in large-scale breeding trials to develop corn hybrids of desired endosperm properties for improved ruminal degradabilities.