Premium
Chemometrics can be applied to mechanical testing data to characterise stem toughness and stiffness in crop plants
Author(s) -
Kemsley EK,
Swinhoe RR,
Smith AC,
McCann MC,
Wang TL
Publication year - 2004
Publication title -
journal of the science of food and agriculture
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.782
H-Index - 142
eISSN - 1097-0010
pISSN - 0022-5142
DOI - 10.1002/jsfa.1731
Subject(s) - flexural strength , flexural modulus , stiffness , univariate , toughness , composite material , multivariate statistics , materials science , modulus , fracture toughness , displacement (psychology) , mathematics , statistics , psychology , psychotherapist
Mechanical tests have been used to assess the engineering properties of pea ( Pisum sativum L) stems. Measurements were made on plants of three different genotypes at four different stages of development and at five defined locations along the stem. The force–displacement curves obtained were used to estimate values of the engineering properties of toughness and flexural modulus, from cutting and flexure mechanical tests respectively. Specimens of all genotypes showed an increase in toughness with age and generally also with stem height. However, there were marked differences in flexural modulus between genotypes. One genotype, known to exhibit a ‘stiff straw’ characteristic, showed a consistent increase in modulus with age and stem height, and at and beyond fruiting had substantially the greatest flexural modulus. The remaining genotypes showed decreasing flexural modulus with age. Chemometric methods were used to analyse sets of complete force–displacement curves, following suitable pre‐processing to allow the application of linear algebra methods. Whereas univariate consideration of the engineering quantities allowed trends to be observed, multivariate analysis of force–distance curves was able to model empirically the genotype differences so that individual specimens could be largely correctly classified. Examination of some of the model coefficients suggested that the ability to discriminate between genotypes is related to structural features of the specimens and that cutting tests in particular are sensitive to the anatomy of the specimen. This is the first time that chemometric methods have been applied to such data and suggests the potential of mechanical tests combined with multivariate analysis to form the basis of a screening system for phenotypic properties of new lines and varieties. Copyright © 2004 Society of Chemical Industry