Non-destructive determination of carbohydrate reserves in leaves of ornamental cuttings by near-infrared spectroscopy (NIRS) as a key indicator for quality assessments

• Published in: Biosystems engineering Vol. 158; pp. 51 - 63
• Main Authors: Lohr, Dieter, Tillmann, Peter, Druege, Uwe, Zerche, Siegfried, Rath, Thomas, Meinken, Elke
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.06.2017
• Subjects: Adventitious rooting; Partial least square regression; Soluble sugars; Starch; Adventitious rooting; Starch; Partial least square regression; Soluble sugars
• ISSN: 1537-5110; 1537-5129
• DOI: 10.1016/j.biosystemseng.2017.03.005

The importance of carbohydrate reserves in leaves for rooting performance of ornamental cuttings is well-known. Especially under environmental conditions unfavourable for photosynthesis, sufficient reserves are indispensable for an undisturbed adventitious root formation and to prevent senescence of leaves during rooting. However, due to time and costs, established methods for carbohydrates analysis are not suitable for implementation in global production chains of ornamentals. Near-infrared spectroscopy (NIRS) might be a valuable alternative. To explore the suitability of this technique, NIR spectra were taken from intact cuttings as well as from upper and lower side of detached leaves of chrysanthemum and pelargonium cuttings and partial least squares (PLS) calibration models were developed for glucose, fructose, sucrose and starch in leaves, which were analysed by a stepwise enzymatic-photometric method. Presumably because of a high percentage of cuttings with very low amounts of glucose, fructose and sucrose, calibration models for single soluble sugars and sum of soluble sugars were poor (RCV2 ≤ 0.5, RPDCV ≤ 1.5), while prediction performance for starch and sum of starch and soluble sugars was quite good (R2 > 0.8, RPD > 2.0, RER > 10). The high number of cuttings with depleted reserves of soluble sugars seems to have been at least partly caused by transportation of cuttings, before NIR analysis, from stock plant facilities in Africa and Latin America to Central Europe. The quite low levels of leaf carbohydrates on delivery at rooting facilities cannot be detected by NIRS properly. Thus, NIRS seems to be more suitable for monitoring of leaf carbohydrates in stock plants to optimise crop management than for assessment of cutting quality before rooting. •Near-infrared spectroscopy is suitable to predict leaf carbohydrates in cuttings.•No sample preparation as drying or chopping is necessary.•Non-normal distribution of calibration data does not affect prediction power.•Significances of validation statistics is arguable for non-normal distributed data.•Geometric mean and standard deviation might be more suitable in such cases.