FIBRE DETERMINATION BY HYDRAULIC PRESSING—WHICH METHOD IS CORRECT?

摘要

Routine analyses of quality components of sugarcane for calibration development of near infra-red spectroscopic applications are enhanced if hydraulic press extraction of juice is used. This results in high-level extraction comparable to that achieved by the first mill in a commercial milling train. A bonus is that fibre also can be estimated with this process. Scepticism has been expressed within the BSES/CSIRO JV breeding team as to the correctness of the method used to determine fibre for the SpectraCane™ calibration. This paper reports on research that establishes the relationship between two methods of determining fibre from hydraulic press data (Berding and Pollock, 1982 (F_(B&P)), Tanimoto, 1964 (F_T)) and fibre determined by a standard method (F_B). Harvest samples were taken from a replicated final assessment trial containing 2 x 72 entries. Two sub-samples from a sample from each entry were processed to yield data for Brix, expressed juice weight, two fibre determinations (F_(B&P) and F_T), and polarimeter reading, all from hydraulic pressing, F_B, and moisture content. Total analytical results were accurate with an average mass balance of 1000.9 g/kg. Use of two sub-samples per sample was satisfactory for all traits, with F_ B and expressed juice having the lowest error ratios of 6 .8 and 6.2, respectively. Mean values for F_B (138.1 g/kg) and Fb&p (137.8 g/kg) were similar while that for F_ T was lower (127.2 g/kg). All fibre determinations were highly correlated, particularly Fb&p and F_T but their regressions against F_B produced significantly different regression coefficients (t140 = 3.129**). Both regressions were skewed, with the latter being more so. The computation of F_(B&P) adjusts the dry plug weight for residual Brix, taking into account hygroscopic water. The F_T computation does not, which results in underestimation of fibre content, The F_T estimates are consistently lower than F_(B&P) and F_B values, and so the F_T methodology must be considered flawed.